Hand_Data_Processing.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\SimJaekyu\\Anaconda3\\lib\\site-packages\\h5py\\__init__.py:36: FutureWarning: Conversion of the second argument of issubdtype from `float` to `np.floating` is deprecated. In future, it will be treated as `np.float64 == np.dtype(float).type`.\n",
      "  from ._conv import register_converters as _register_converters\n"
     ]
    }
   ],
   "source": [
    "import cv2\n",
    "import ast\n",
    "import os\n",
    "import json\n",
    "import numpy as np\n",
    "\n",
    "\n",
    "import tensorflow as tf\n",
    "import math\n",
    "import collections\n",
    "import sys\n",
    "import matplotlib.pyplot as plt\n",
    "from matplotlib.image import imread\n",
    "import matplotlib\n",
    "import os.path"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "hand_train_data = os.listdir(\"./dataset/Hands/hand_labels/manual_train/\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "3824"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(hand_train_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "hands_json_data = sorted(f for f in hand_train_data if f.endswith('.json'))\n",
    "hands_img_data = sorted(f for f in hand_train_data if not f.endswith('.json'))\n",
    "\n",
    "_hands_json_data = np.copy(hands_json_data)\n",
    "_hands_img_data = np.copy(hands_img_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1912\n",
      "1912\n",
      "000015774_01_l.json\n",
      "000015774_01_l.jpg\n"
     ]
    }
   ],
   "source": [
    "print(len(hands_img_data))\n",
    "print(len(hands_json_data))\n",
    "print(hands_json_data[0])\n",
    "print(_hands_img_data[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "./dataset/Hands/hand_labels/manual_train/000015774_01_l.jpg\n",
      "./dataset/Hands/hand_labels/manual_train/000015774_01_l.json\n"
     ]
    }
   ],
   "source": [
    "for i in range(len(hands_img_data)):\n",
    "    hands_img_data[i] = \"./dataset/Hands/hand_labels/manual_train/\" + hands_img_data[i]\n",
    "    \n",
    "for i in range(len(hands_json_data)):\n",
    "    hands_json_data[i] = \"./dataset/Hands/hand_labels/manual_train/\" + hands_json_data[i]\n",
    "    \n",
    "print(hands_img_data[0])\n",
    "print(hands_json_data[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'000015774_01_l'"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "_hands_img_data[0][:-4]\n",
    "_hands_json_data[0][:-5]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1912"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(hands_img_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1912"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(hands_json_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "#img size\n",
    "_tmp = []\n",
    "for index in range(len(hands_img_data)):\n",
    "    if(_hands_json_data[index][:-5] == _hands_img_data[index][:-4]):\n",
    "        img_data = cv2.imread(hands_img_data[index])\n",
    "        img_height = np.shape(img_data)[0]\n",
    "        img_width = np.shape(img_data)[1]\n",
    "        #joint data load\n",
    "        f = open(hands_json_data[index])\n",
    "        s = f.readlines()\n",
    "        hands_index_s = s[0].index(\"hand_pts\")\n",
    "        hands_index_f = s[0].index(\"head_box\")\n",
    "        hands_pts_data = s[0][hands_index_s+11 : hands_index_f-3]\n",
    "\n",
    "        #joint data 변환\n",
    "        alpha = ast.literal_eval(hands_pts_data)\n",
    "        data = dict()\n",
    "        for i in range(len(alpha)):\n",
    "            x_point = int(alpha[i][0])#int((alpha[i][0] / img_width) * 44)\n",
    "            y_point = int(alpha[i][1])#int((alpha[i][1] / img_height) * 44)\n",
    "            data[\"%d\"%i] = [x_point, y_point]\n",
    "        _tmp.append(data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1912,)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{'0': [847, 522],\n",
       " '1': [837, 525],\n",
       " '2': [828, 528],\n",
       " '3': [813, 548],\n",
       " '4': [817, 562],\n",
       " '5': [838, 524],\n",
       " '6': [836, 552],\n",
       " '7': [837, 560],\n",
       " '8': [840, 570],\n",
       " '9': [853, 526],\n",
       " '10': [848, 551],\n",
       " '11': [846, 561],\n",
       " '12': [844, 569],\n",
       " '13': [862, 530],\n",
       " '14': [860, 549],\n",
       " '15': [857, 556],\n",
       " '16': [853, 563],\n",
       " '17': [870, 534],\n",
       " '18': [869, 552],\n",
       " '19': [865, 557],\n",
       " '20': [862, 561]}"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(np.shape(_tmp))\n",
    "(_tmp[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[847, 522]"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "_tmp[0]['0']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'./dataset/Hands/hand_labels/manual_train/000015774_01_l.jpg'"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "hands_img_data[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "img_data = cv2.imread(hands_img_data[0])\n",
    "plt.imshow(img_data)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "해당 셀은 손 부분만 크롭하여 이미지로 저장하는 셀. \n",
    "for index in range(len(hands_img_data)):\n",
    "    img_data = cv2.imread(hands_img_data[index])\n",
    "    x_max = 0\n",
    "    y_max = 0\n",
    "    x_min = 9999\n",
    "    y_min = 9999\n",
    "    for i in range(len(_tmp[index])):\n",
    "        if(x_max < _tmp[index][\"%d\"%i][0]):\n",
    "            x_max = _tmp[index][\"%d\"%i][0]\n",
    "        if(y_max < _tmp[index][\"%d\"%i][1]):\n",
    "            y_max = _tmp[index][\"%d\"%i][1]\n",
    "        if(x_min > _tmp[index][\"%d\"%i][0]):\n",
    "            x_min = _tmp[index][\"%d\"%i][0]\n",
    "        if(y_min > _tmp[index][\"%d\"%i][1]):\n",
    "            y_min = _tmp[index][\"%d\"%i][1]\n",
    "    x_width = x_max-x_min\n",
    "    y_width = y_max-y_min\n",
    "\n",
    "    length = x_width if x_width > y_width else y_width\n",
    "\n",
    "    length\n",
    "    x_center = ((x_max + x_min) // 2)\n",
    "    y_center = ((y_max + y_min) // 2)\n",
    "    print(x_center, y_center)\n",
    "    crop_x_min = x_center - int(1.1 * length)\n",
    "    crop_x_max = x_center + int(1.1 * length)\n",
    "    crop_y_min = y_center - int(1.1 * length)\n",
    "    crop_y_max = y_center + int(1.1 * length)\n",
    "    if(crop_x_min < 0):\n",
    "        crop_x_min = 0\n",
    "    if(crop_x_max >= np.shape(img_data)[1]):\n",
    "        crop_x_max = np.shape(img_data)[1]\n",
    "    if(crop_y_min < 0):\n",
    "        crop_y_min = 0\n",
    "    if(crop_y_max >= np.shape(img_data)[0]):\n",
    "        crop_y_max = np.shape(img_data)[0]\n",
    "    _map = np.zeros((crop_y_max - crop_y_min, crop_x_max - crop_x_min, 3), dtype=np.int16)\n",
    "    \n",
    "    for i in range(crop_x_max-crop_x_min):\n",
    "        for j in range(crop_y_max-crop_y_min):\n",
    "            #try:\n",
    "            _map[j][i] = img_data[crop_y_min + j][crop_x_min + i]\n",
    "    _map = np.transpose(_map, [1, 0, 2])\n",
    "    _map = _map[:, :, [2, 0, 1]]\n",
    "    f_name_index = hands_img_data[index].index(\"train/\")+6\n",
    "    f_name = hands_img_data[index][f_name_index : ]\n",
    "    _path = \"./dataset/Hands/hand_labels/Cropped_Image/\" + f_name\n",
    "    cv2.imwrite(_path,_map[:, :, [2, 1, 0]])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "#해당 부분은 윗 셀에서 저장한 crop된 셀이 옆으로 저장되어서 다시 transpose하고, 색 체계도 바꾸는 셀\n",
    "file_list = os.listdir(\"./dataset/Hands/hand_labels/Cropped_Image/\")\n",
    "tmp = []\n",
    "for i in range(len(file_list)):\n",
    "    tmp.append(\"./dataset/Hands/hand_labels/Cropped_Image/\" + file_list[i])\n",
    "#file_list = tmp\n",
    "for i in range(len(tmp)):\n",
    "    #print(tmp[i])\n",
    "    img_data = cv2.imread(tmp[i])\n",
    "    #print(np.shape(img_data))\n",
    "    #plt.imshow(img_data)\n",
    "    #plt.show()\n",
    "    img_data = np.transpose(img_data, [1, 0, 2])\n",
    "    img_data = cv2.resize(img_data, (356, 356))\n",
    "    f_name_index = tmp[i].index(\"Image/\")+6\n",
    "    f_name = tmp[i][f_name_index : ]\n",
    "    _path = \"./dataset/Hands/hand_labels/Cropped_Resized_Image/\" + f_name\n",
    "    cv2.imwrite(_path, img_data[:, :, :])\n",
    "    #plt.imshow(img_data[:, :, [2, 1, 0]])\n",
    "    #plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'0': [847, 522],\n",
       " '1': [837, 525],\n",
       " '2': [828, 528],\n",
       " '3': [813, 548],\n",
       " '4': [817, 562],\n",
       " '5': [838, 524],\n",
       " '6': [836, 552],\n",
       " '7': [837, 560],\n",
       " '8': [840, 570],\n",
       " '9': [853, 526],\n",
       " '10': [848, 551],\n",
       " '11': [846, 561],\n",
       " '12': [844, 569],\n",
       " '13': [862, 530],\n",
       " '14': [860, 549],\n",
       " '15': [857, 556],\n",
       " '16': [853, 563],\n",
       " '17': [870, 534],\n",
       " '18': [869, 552],\n",
       " '19': [865, 557],\n",
       " '20': [862, 561]}"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "_tmp[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(480, 654, 3)\n",
      "(480, 654, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(480, 848, 3)\n",
      "(480, 848, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1278, 3)\n",
      "(720, 1278, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1440, 3)\n",
      "(1080, 1440, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1916, 3)\n",
      "(1080, 1916, 3)\n",
      "(1080, 1916, 3)\n",
      "(1080, 1916, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(480, 854, 3)\n",
      "(480, 854, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1440, 3)\n",
      "(1080, 1440, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(480, 654, 3)\n",
      "(480, 654, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 960, 3)\n",
      "(720, 960, 3)\n",
      "(720, 960, 3)\n",
      "(720, 960, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1916, 3)\n",
      "(1080, 1916, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1440, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1916, 3)\n",
      "(1080, 1916, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1440, 3)\n",
      "(1080, 1440, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1906, 3)\n",
      "(1080, 1906, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1440, 3)\n",
      "(1080, 1440, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1906, 3)\n",
      "(1080, 1906, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(480, 854, 3)\n",
      "(480, 854, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1906, 3)\n",
      "(1080, 1906, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(480, 854, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1440, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1440, 3)\n",
      "(1080, 1440, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 960, 3)\n",
      "(720, 960, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1906, 3)\n",
      "(1080, 1906, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 960, 3)\n",
      "(720, 960, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(1080, 1920, 3)\n",
      "(720, 1280, 3)\n",
      "(720, 1280, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(576, 720, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n",
      "(300, 400, 3)\n"
     ]
    }
   ],
   "source": [
    "for index in range(len(hands_img_data)):\n",
    "    #print(hands_img_data[0])\n",
    "    img_data_ = cv2.imread(hands_img_data[index])\n",
    "    #print(np.shape(img_data_))\n",
    "    x_max = 0\n",
    "    y_max = 0\n",
    "    x_min = 9999\n",
    "    y_min = 9999\n",
    "    for i in range(len(buffer)):\n",
    "        if(x_max < _tmp[index][\"%d\"%i][0]):\n",
    "            x_max = _tmp[index][\"%d\"%i][0]\n",
    "        if(y_max < _tmp[index][\"%d\"%i][1]):\n",
    "            y_max = _tmp[index][\"%d\"%i][1]\n",
    "        if(x_min > _tmp[index][\"%d\"%i][0]):\n",
    "            x_min = _tmp[index][\"%d\"%i][0]\n",
    "        if(y_min > _tmp[index][\"%d\"%i][1]):\n",
    "            y_min = _tmp[index][\"%d\"%i][1]\n",
    "    x_width = x_max-x_min\n",
    "    y_width = y_max-y_min\n",
    "\n",
    "    length = x_width if x_width > y_width else y_width\n",
    "\n",
    "    length\n",
    "    x_center = ((x_max + x_min) // 2)\n",
    "    y_center = ((y_max + y_min) // 2)\n",
    "    #print(x_center, y_center)\n",
    "    crop_x_min = x_center - int(1.1 * length)\n",
    "    crop_x_max = x_center + int(1.1 * length)\n",
    "    crop_y_min = y_center - int(1.1 * length)\n",
    "    crop_y_max = y_center + int(1.1 * length)\n",
    "\n",
    "    size = length * 2.2\n",
    "\n",
    "    for i in range(len(buffer)):\n",
    "        _tmp[index][\"%d\"%i][0] = _tmp[index][\"%d\"%i][0] - crop_x_min\n",
    "        _tmp[index][\"%d\"%i][1] = _tmp[index][\"%d\"%i][1] - crop_y_min\n",
    "\n",
    "    for i in range(len(buffer)):\n",
    "        _tmp[index][\"%d\"%i][0] = int((_tmp[index][\"%d\"%i][0] / size) * 44)\n",
    "        _tmp[index][\"%d\"%i][1] = int((_tmp[index][\"%d\"%i][1] / size) * 44)\n",
    "    _tmp[index]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [],
   "source": [
    "#np.save(\"./dataset/Hands/hand_labels/Annotation/cropped_resized_hands_joint_data.npy\", _tmp)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'000015774_01_l.jpg'"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#_hands_img_data[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "#path_data = []\n",
    "#for i in range(len(_hands_img_data)):\n",
    "#    path_data.append(\"./dataset/Hands/hand_labels/Cropped_Resized_Image/\" + _hands_img_data[i])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "#np.save(\"./dataset/Hands/hand_labels/Annotation/cropped_resized_hands_img_list.npy\", path_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "#np.save('./dataset/Hands/hand_labels/Annotation/hands_joint_data.npy', _tmp) # x_save.npy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "#img_list = []\n",
    "#for i in range(len(hands_img_data)):\n",
    "#    img_list.append(\"./dataset/Hands/hand_labels/Resized_Image/\" + _hands_img_data[i])\n",
    "#np.save('./dataset/Hands/hand_labels/Annotation/resized_hands_img_list.npy', img_list) # x_save.npy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "#joint_data = np.load('./dataset/Hands/hand_labels/Annotation/hands_joint_data.npy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "#img_list = np.load('./dataset/Hands/hand_labels/Annotation/resized_hands_img_list.npy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "21"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#len(joint_data[-5])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "def make_heatmap(batch_anno_data, width = 44, height = 44, num_of_maps = 22):\n",
    "    batch_size = len(batch_anno_data)\n",
    "    width = 44\n",
    "    height = 44\n",
    "    num_of_maps = 22\n",
    "    output = np.zeros((batch_size, width, height, num_of_maps))\n",
    "    for index, joint_data in enumerate(batch_anno_data):\n",
    "\n",
    "        heatmap = np.zeros((width, height, num_of_maps), np.int16)#batch 일단 뺌\n",
    "\n",
    "        for joints in joint_data:\n",
    "            #print(joints)\n",
    "            buffer = list(joints.items())\n",
    "            key_buffer = joints.keys()\n",
    "\n",
    "            for i in range(len(buffer)):\n",
    "                buffer[i] = list(buffer[i])\n",
    "                buffer[i][0] = int(buffer[i][0])\n",
    "            bubble_sort(buffer)\n",
    "\n",
    "            idx = 0\n",
    "            for j in range(22):\n",
    "\n",
    "                #print(j)\n",
    "                if('%d' %j in key_buffer):\n",
    "                    center_x = buffer[idx][1][0]\n",
    "                    center_y = buffer[idx][1][1]\n",
    "                    #joint = buffer[idx][1]\n",
    "                    joint = [center_y, center_x]\n",
    "                    idx = idx + 1\n",
    "                    _put_heatmap_on_plane(heatmap, plane_idx = j, joint = joint, sigma = 3, height = height, width = width, stride = 1)\n",
    "                else:\n",
    "                    pass\n",
    "            idx = 0\n",
    "        heatmap[:, :, -1] = np.clip(1.0 - np.amax(heatmap, axis=2), 0.0, 1.0)\n",
    "        output[index] = heatmap\n",
    "    return output\n",
    "\n",
    "def make_paf_field(batch_joint_data, width = 44, height = 44, num_of_maps = 22):\n",
    "    batch_size = len(batch_joint_data)\n",
    "    output1 = np.zeros((batch_size, width, height, num_of_maps*2))\n",
    "    output2 = np.zeros((batch_size, width, height, num_of_maps))\n",
    "    for index, joint_data in enumerate(batch_joint_data):\n",
    "        joint_pairs = list(zip(\n",
    "            [0, 1, 2, 3, 0, 5, 6, 7, 0, 10, 11,  0, 13, 14, 15,  0, 17, 18, 19],\n",
    "            [1, 2, 3, 4, 5, 6, 7, 8, 9, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20]))\n",
    "        #make vector map\n",
    "        width = 44\n",
    "        height = 44\n",
    "        num_of_maps = 22\n",
    "        vectormap = np.zeros((width, height, num_of_maps*2), dtype=np.float32)#batch 일단 뺌\n",
    "        countmap = np.zeros((width, height, num_of_maps), np.int16)#batch 일단 뺌\n",
    "\n",
    "        for joints in joint_data:\n",
    "            key = (joints.keys())\n",
    "            for plane_idx, (j_idx1, j_idx2) in enumerate(joint_pairs):\n",
    "                if(('%d' %j_idx1 in key) and ('%d' %j_idx2 in key)):\n",
    "\n",
    "                    center_from = joints['%d'%j_idx1]\n",
    "                    center_to = joints['%d'%j_idx2]\n",
    "\n",
    "                    if not center_from or not center_to:\n",
    "                        continue\n",
    "                    _put_paf_on_plane(vectormap=vectormap, countmap=countmap, plane_idx=plane_idx, center_from=center_from, center_to=center_to, threshold=1, height=88, width=88, stride = 1)\n",
    "\n",
    "        nonzeros = np.nonzero(countmap)\n",
    "\n",
    "\n",
    "        for x, y, p in zip(nonzeros[0], nonzeros[1], nonzeros[2]):\n",
    "            if countmap[x][y][p] <= 0:\n",
    "                continue\n",
    "            vectormap[x][y][p*2+0] /= countmap[x][y][p]\n",
    "            vectormap[x][y][p*2+1] /= countmap[x][y][p]\n",
    "\n",
    "        output1[index] = vectormap.astype(np.float32)\n",
    "        output2[index] = countmap\n",
    "    return output1, output2 #output1 -> vectormap, output2 -> countmap\n",
    "\n",
    "def _put_heatmap_on_plane(heatmap, plane_idx, joint, sigma, height, width, stride):\n",
    "    start = stride / 2.0 - 0.5\n",
    "\n",
    "    center_x, center_y = joint\n",
    "\n",
    "    for g_y in range(height):\n",
    "        for g_x in range(width):\n",
    "            x = start + g_x * stride\n",
    "            y = start + g_y * stride\n",
    "            d2 = (x-center_x) * (x-center_x) + (y-center_y) * (y-center_y)\n",
    "            exponent = d2 / 2.0 / sigma / sigma\n",
    "            if exponent > 4.6052:\n",
    "                continue\n",
    "\n",
    "            heatmap[g_y, g_x, plane_idx] += math.exp(-exponent)\n",
    "            if heatmap[g_y, g_x, plane_idx] > 1.0:\n",
    "                heatmap[g_y, g_x, plane_idx] = 1.0\n",
    "                \n",
    "                \n",
    "def _put_paf_on_plane(vectormap, countmap, plane_idx, center_from, center_to, threshold, height, width, stride):\n",
    "    center_from = (center_from[0] // stride, center_from[1] // stride)\n",
    "    center_to = (center_to[0] // stride, center_to[1] // stride)\n",
    "\n",
    "    vec_x = center_to[0] - center_from[0]\n",
    "    vec_y = center_to[1] - center_from[1]\n",
    "\n",
    "    min_x = max(0, int(min(center_from[0], center_to[0]) - threshold))\n",
    "    min_y = max(0, int(min(center_from[1], center_to[1]) - threshold))\n",
    "\n",
    "    max_x = min(width, int(max(center_from[0], center_to[0]) + threshold))\n",
    "    max_y = min(height, int(max(center_from[1], center_to[1]) + threshold))\n",
    "\n",
    "    norm = math.sqrt(vec_x ** 2 + vec_y ** 2)\n",
    "    if norm < 1e-8: #1e-8 이하는 0으로 인식되서 0으로 나눌수 없다는 에러 발생. 따라서 return처리 해줌\n",
    "        return\n",
    "\n",
    "    vec_x /= norm\n",
    "    vec_y /= norm\n",
    "    \n",
    "    for x in range(min_x, max_x):\n",
    "        for y in range(min_y, max_y):\n",
    "            bec_x = x - center_from[0]\n",
    "            bec_y = y - center_from[1]\n",
    "            dist = abs(bec_x * vec_y - bec_y * vec_x)\n",
    "\n",
    "            if dist > threshold:\n",
    "                continue\n",
    "\n",
    "            countmap[x][y][plane_idx] = countmap[x][y][plane_idx] + 1\n",
    "\n",
    "            vectormap[x][y][plane_idx*2+0] = vec_x\n",
    "            vectormap[x][y][plane_idx*2+1] = vec_y\n",
    "def bubble_sort(L):\n",
    "    for i in range(len(L)-1):\n",
    "        for j in range(len(L)-1):\n",
    "            if L[j] > L[j+1]:\n",
    "                temp = L[j+1]\n",
    "                L[j+1] = L[j]\n",
    "                L[j] = temp      "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "#plt.imshow(img_data[:,:,[2,1,0]])\n",
    "#plt.imshow(tmp[:,:], alpha=.7)#alpha -> 이미지 얼마나 선명한가. heatmap_avg[:,:,0] -> 1:r_knee\n",
    "#fig = matplotlib.pyplot.gcf()\n",
    "#cax = matplotlib.pyplot.gca()\n",
    "#fig.set_size_inches(20, 20)\n",
    "#fig.subplots_adjust(right=0.93)\n",
    "#cbar_ax = fig.add_axes([0.95, 0.15, 0.01, 0.7])\n",
    "#_ = fig.colorbar(oriImg, cax=cbar_ax)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABLMAAARiCAYAAABf4LXoAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvhp/UCwAAIABJREFUeJzs3V2sped51+H/XU+GNEOCEzIOxuPIPpiSWJWawJZlyQekCUFOQLYP0shWAYOszknNhxoBLkIBAkgtBw0gGdCIRJkiGsdqKR5FI9zIJCqgJnhM2hDbWBlMiEe2MiZ1Qhk+ptPeHOzlZHvP3t57ZvZH76zrkkZ7Pc961qz7+Kf3fVd1dwAAAABggh/Y7wEAAAAAYLvELAAAAADGELMAAAAAGEPMAgAAAGAMMQsAAACAMcQsAAAAAMYQswAAAAAYQ8wCAAAAYAwxCwAAAIAxxCwAAAAAxjiwl1/2hoOH+to3vHkvvxIAAABgz5z7zjcvXOzf/n37Pcf3sz2NWde+4c25/48/sJdfCQAAALBnfu7k3/8/+z3D97urus2wqu6oqmer6kxVPbhTQwEAAADARq44ZlXVNUkeSvKBJLckubeqbtmpwQAAAABgvau5MuvWJGe6+7nuvpDk4SR37cxYAAAAAHCpq4lZNyR5fs367GIPAAAAAHbF1cSs2mCvLzlUdayqTlfV6fMXzl/F1wEAAACw7K4mZp1NcuOa9ZEkL6w/1N3Hu3ulu1cOHTx0FV8HAAAAwLK7mpj1RJKjVXVzVR1Mck+SkzszFgAAAABc6sCVfrC7L1bVA0keS3JNkk9291M7NhkAAAAArHPFMStJuvtUklM7NAsAAAAAvKaruc0QAAAAAPaUmAUAAADAGGIWAAAAAGOIWQAAAACMIWYBAAAAMIaYBQAAAMAYYhYAAAAAY4hZAAAAAIwhZgEAAAAwhpgFAAAAwBhiFgAAAABjiFkAAAAAjCFmAQAAADCGmAUAAADAGGIWAAAAAGOIWQAAAACMIWYBAAAAMIaYBQAAAMAYYhYAAAAAY4hZAAAAAIwhZgEAAAAwhpgFAAAAwBhiFgAAAABjiFkAAAAAjCFmAQAAADCGmAUAAADAGGIWAAAAAGOIWQAAAACMIWYBAAAALKGq+mRVnauqr27yflXVP66qM1X1lar6o3s940bELAAAAIDl9Kkkd7zG+x9IcnTx71iSf7oHM21JzAIAAABYQt39q0l+8zWO3JXk53vVF5NcW1XX7810mxOzAAAAANjIDUmeX7M+u9jbVwf2ewAAAAAAfk+qDfZ6w4NVx7J6K2IOHTr0x97xjndc9pc9+eST/6O7D291TswCAAAAYCNnk9y4Zn0kyQsbHezu40mOJ8nKykqfPn36sr+sqv77ds65zRAAAACAjZxM8ucWv2p4W5LvdPeL+z2UK7MAAAAAllBVfTrJe5K8tarOJvlbSV6XJN39z5KcSvLBJGeS/O8kf2F/Jn01MQsAAABgCXX3vVu830l+co/G2Ta3GQIAAAAwhpgFAAAAwBhiFgAAAABjiFkAAAAAjCFmAQAAADCGmAUAAADAGGIWAAAAAGOIWQAAAACMIWYBAAAAMIaYBQAAAMAYYhYAAAAAY4hZAAAAAIwhZgEAAAAwhpgFAAAAwBhiFgAAAABjiFkAAAAAjCFmAQAAADCGmAUAAADAGGIWAAAAAGOIWQAAAACMIWYBAAAAMIaYBQAAAMAYYhYAAAAAY4hZAAAAAIwhZgEAAAAwhpgFAAAAwBhiFgAAAABjiFkAAAAAjCFmAQAAADCGmAUAAADAGGIWAAAAAGOIWQAAAACMIWYBAAAAMIaYBQAAAMAYYhYAAAAAY4hZAAAAAIwhZgEAAAAwhpgFAAAAwBhiFgAAAABjiFkAAAAAjCFmAQAAADCGmAUAAADAGGIWAAAAAGOIWQAAAACMIWYBAAAAMIaYBQAAAMAYYhYAAAAAY4hZAAAAAIwhZgEAAAAwhpgFAAAAwBhiFgAAAABjiFkAAAAAjCFmAQAAADCGmAUAAADAGGIWAAAAAGOIWQAAAACMIWYBAAAAMIaYBQAAAMAYYhYAAAAAY4hZAAAAAIwhZgEAAAAwhpgFAAAAwBhiFgAAAABjiFkAAAAAjCFmAQAAADCGmAUAAADAGGIWAAAAAGOIWQAAAACMIWYBAAAAMIaYBQAAAMAYYhYAAAAAY4hZAAAAAIwhZgEAAAAwhpgFAAAAwBhiFgAAAABjiFkAAAAAjCFmAQAAADDGljGrqj5ZVeeq6qtr9t5SVZ+rqq8t/r55d8cEAAAAgO1dmfWpJHes23swyePdfTTJ44s1AAAAAOyqLWNWd/9qkt9ct31XkhOL1yeS3L3DcwEAAADAJa70mVlv6+4Xk2Tx97qdGwkAAAAANrbrD4CvqmNVdbqqTp+/cH63vw4AAACA72NXGrO+WVXXJ8ni77nNDnb38e5e6e6VQwcPXeHXAQAAAMCVx6yTSe5bvL4vyaM7Mw4AAAAAbG7LmFVVn07ya0n+SFWdrar7k/xMkvdX1deSvH+xBgAAAIBddWCrA9197yZvvW+HZwEAAACA17TrD4AHAAAAgJ0iZgEAAAAwhpgFAAAAwBhiFgAAAABjiFkAAAAAjCFmAQAAADCGmAUAAADAGGIWAAAAAGOIWQAAAACMIWYBAAAAMIaYBQAAAMAYYhYAAAAAY4hZAAAAAIwhZgEAAAAwhpgFAAAAwBhiFgAAAABjiFkAAAAAjCFmAQAAADCGmAUAAADAGGIWAAAAAGOIWQAAAACMIWYBAAAAMIaYBQAAAMAYYhYAAAAAY4hZAAAAAIwhZgEAAAAwhpgFAAAAwBhiFgAAAABjiFkAAAAAjCFmAQAAADCGmAUAAADAGGIWAAAAAGOIWQAAAACMIWYBAAAAMIaYBQAAAMAYYhYAAAAAY4hZAAAAAIwhZgEAAAAwhpgFAAAAwBhiFgAAAABjiFkAAAAAjCFmAQAAADCGmAUAAADAGGIWAAAAAGOIWQAAAACMIWYBAAAAMIaYBQAAAMAYYhYAAAAAY4hZAAAAAIwhZgEAAAAwhpgFAAAAwBhiFgAAAABjiFkAAAAAjCFmAQAAADCGmAUAAADAGGIWAAAAAGOIWQAAAACMIWYBAAAAMIaYBQAAAMAYYhYAAAAAY4hZAAAAAIwhZgEAAAAwhpgFAAAAwBhiFgAAAABjiFkAAAAAjCFmAQAAADCGmAUAAADAGGIWAAAAAGOIWQAAAACMIWYBAAAAMIaYBQAAAMAYYhYAAAAAY4hZAAAAAIwhZgEAAAAwhpgFAAAAwBhiFgAAAABjiFkAAAAAjCFmAQAAADCGmAUAAADAGGIWAAAAAGOIWQAAAACMIWYBAAAAMIaYBQAAAMAYYhYAAAAAY4hZAAAAAIwhZgEAAAAwhpgFAAAAwBhiFgAAAABjiFkAAAAAjCFmAQAAADCGmAUAAADAGGIWAAAAAGOIWQAAAACMIWYBAAAAMIaYBQAAAMAYYhYAAAAAY4hZAAAAAIwhZgEAAAAwhpgFAAAAwBhiFgAAAABjiFkAAAAAjCFmAQAAADCGmAUAAADAGGIWAAAAAGOIWQAAAACMIWYBAAAAMIaYBQAAAMAYYhYAAADAkqqqO6rq2ao6U1UPbvD+26vq81X15ar6SlV9cD/mXEvMAgAAAFhCVXVNkoeSfCDJLUnurapb1h37m0ke6e53J7knyT/Z2ykvJWYBAAAALKdbk5zp7ue6+0KSh5Pcte5MJ3nT4vUfSPLCHs63ITELAAAAYDndkOT5Neuzi721/naSP1NVZ5OcSvIXN/qPqupYVZ2uqtMvvfTSbsz6XWIWAAAAwHKqDfZ63freJJ/q7iNJPpjkX1TVJT2pu49390p3rxw+fHgXRv0eMQsAAABgOZ1NcuOa9ZFcehvh/UkeSZLu/rUkr0/y1j2ZbhNiFgAAAMByeiLJ0aq6uaoOZvUB7yfXnflGkvclSVW9M6sxa3fvI9yCmAUAAACwhLr7YpIHkjyW5Jms/mrhU1X1saq6c3HsI0l+oqp+I8mnk/z57l5/K+KeOrDVgaq6McnPJ/lDSX43yfHu/kdV9ZYkn0lyU5KvJ/lwd7+8e6MCAAAAsJO6+1RWH+y+du+ja14/neT2vZ7rtWznyqyLST7S3e9McluSn6yqW5I8mOTx7j6a5PHFGgAAAAB2zZYxq7tf7O7/tHj9W1m97OyGJHclObE4diLJ3bs1JAAAAAAkl/nMrKq6Kcm7k3wpydu6+8VkNXgluW6nhwMAAACAtbYds6rq9yf5pSR/pbv/52V87lhVna6q0+cvnL+SGQEAAAAgyTZjVlW9Lqsh6192979abH+zqq5fvH99knMbfba7j3f3SnevHDp4aCdmBgAAAGBJbRmzqqqSfCLJM939c2veOpnkvsXr+5I8uvPjAQAAAMD3HNjGmduT/Nkk/7mqfn2x9zeS/EySR6rq/iTfSPJjuzMiAAAAAKzaMmZ1979PUpu8/b6dHQcAAAAANndZv2YIAAAAAPtJzAIAAABgDDELAAAAgDHELAAAAADGELMAAAAAGEPMAgAAAGAMMQsAAACAMcQsAAAAAMYQswAAAAAYQ8wCAAAAYAwxCwAAAIAxxCwAAAAAxhCzAAAAABhDzAIAAABgDDELAAAAgDHELAAAAADGELMAAAAAGEPMAgAAAGAMMQsAAACAMcQsAAAAAMYQswAAAAAYQ8wCAAAAYAwxCwAAAIAxxCwAAAAAxhCzAAAAABhDzAIAAABgDDELAAAAgDHELAAAAADGELMAAAAAGEPMAgAAAGAMMQsAAACAMcQsAAAAAMYQswAAAAAYQ8wCAAAAYAwxCwAAAIAxxCwAAAAAxhCzAAAAABhDzAIAAABgDDELAAAAgDHELAAAAADGELMAAAAAGEPMAgAAAGAMMQsAAACAMcQsAAAAAMYQswAAAAAYQ8wCAAAAYAwxCwAAAIAxxCwAAAAAxhCzAAAAABhDzAIAAABgDDELAAAAgDHELAAAAADGELMAAAAAGEPMAgAAAGAMMQsAAACAMcQsAAAAAMYQswAAAAAYQ8wCAAAAYAwxCwAAAIAxxCwAAAAAxhCzAAAAABhDzAIAAABgDDELAAAAgDHELAAAAADGELMAAAAAGEPMAgAAAGAMMQsAAACAMcQsAAAAAMYQswAAAAAYQ8wCAAAAYAwxCwAAAIAxxCwAAAAAxhCzAAAAABhDzAIAAABgDDELAAAAgDHELAAAAADGELMAAAAAGEPMAgAAAGAMMQsAAACAMcQsAAAAAMYQswAAAAAYQ8wCAAAAYAwxCwAAAIAxxCwAAAAAxhCzAAAAABhDzAIAAABgDDELAAAAgDHELAAAAADGELMAAAAAGEPMAgAAAGAMMQsAAACAMcQsAAAAAMYQswAAAAAYQ8wCAAAAYAwxCwAAAIAxxCwAAAAAxhCzAAAAABhDzAIAAABgDDELAAAAgDHELAAAAADGELMAAAAAGEPMAgAAAGAMMQsAAACAMcQsAAAAAMYQswAAAAAYQ8wCAAAAYAwxCwAAAIAxxCwAAAAAxhCzAAAAABhDzAIAAABgDDELAAAAgDHELAAAAADGELMAAAAAGEPMAgAAAGAMMQsAAACAMbaMWVX1+qr6j1X1G1X1VFX9ncX+zVX1par6WlV9pqoO7v64AAAAACyz7VyZ9f+SvLe7fyTJu5LcUVW3JfnZJB/v7qNJXk5y/+6NCQAAAADbiFm96n8tlq9b/Osk703yi4v9E0nu3pUJAQAAAGBhW8/MqqprqurXk5xL8rkk/zXJt7v74uLI2SQ3bPLZY1V1uqpOn79wfidmBgAAAGBJbStmdffvdPe7khxJcmuSd250bJPPHu/ule5eOXTw0JVPCgAAAMDSu6xfM+zubyf5QpLbklxbVQcWbx1J8sLOjgYAAAAAr7adXzM8XFXXLl7/YJI/keSZJJ9P8qHFsfuSPLpbQwIAAABAkhzY+kiuT3Kiqq7Javx6pLs/W1VPJ3m4qv5eki8n+cQuzgkAAAAAW8es7v5KkndvsP9cVp+fBQAAAAB74rKemQUAAAAA+0nMAgAAAGAMMQsAAACAMcQsAAAAAMYQswAAAAAYQ8wCAAAAYAwxCwAAAIAxxCwAAAAAxhCzAAAAABhDzAIAAABgDDELAAAAgDHELAAAAADGELMAAAAAGEPMAgAAAGAMMQsAAACAMcQsAAAAAMYQswAAAAAYQ8wCAAAAYAwxCwAAAIAxxCwAAAAAxhCzAAAAABhDzAIAAABgDDELAAAAgDHELAAAAADGELMAAAAAGEPMAgAAAGAMMQsAAACAMcQsAAAAAMYQswAAAAAYQ8wCAAAAYAwxCwAAAIAxxCwAAAAAxhCzAAAAABhDzAIAAABgDDELAAAAgDHELAAAAADGELMAAAAAGEPMAgAAAGAMMQsAAACAMcQsAAAAAMYQswAAAAAYQ8wCAAAAYAwxCwAAAIAxxCwAAAAAxhCzAAAAABhDzAIAAABgDDELAAAAgDHELAAAAADGELMAAAAAGEPMAgAAAGAMMQsAAACAMcQsAAAAAMYQswAAAAAYQ8wCAAAAYAwxCwAAAIAxxCwAAAAAxhCzAAAAABhDzAIAAABgDDELAAAAgDHELAAAAADGELMAAAAAllRV3VFVz1bVmap6cJMzH66qp6vqqar6hb2ecb0D+z0AAAAAAHuvqq5J8lCS9yc5m+SJqjrZ3U+vOXM0yU8nub27X66q6/Zn2u9xZRYAAADAcro1yZnufq67LyR5OMld6878RJKHuvvlJOnuc3s84yXELAAAAIDldEOS59eszy721vqhJD9UVf+hqr5YVXfs2XSbcJshAAAAwHKqDfZ63fpAkqNJ3pPkSJJ/V1U/3N3fftV/VHUsybEkefvb377zk67hyiwAAACA5XQ2yY1r1keSvLDBmUe7+7e7+78leTarcetVuvt4d69098rhw4d3beBEzAIAAABYVk8kOVpVN1fVwST3JDm57sy/TvKjSVJVb83qbYfP7emU64hZAAAAAEuouy8meSDJY0meSfJIdz9VVR+rqjsXxx5L8q2qejrJ55P81e7+1v5MvMozswAAAACWVHefSnJq3d5H17zuJD+1+Pd7giuzAAAAABhDzAIAAABgDDELAAAAgDHELAAAAADGELMAAAAAGEPMAgAAAGAMMQsAAACAMcQsAAAAAMYQswAAAAAYQ8wCAAAAYAwxCwAAAIAxxCwAAAAAxhCzAAAAABhDzAIAAABgDDELAAAAgDHELAAAAADGELMAAAAAGEPMAgAAAGAMMQsAAACAMcQsAAAAAMYQswAAAAAYQ8wCAAAAYAwxCwAAAIAxxCwAAAAAxhCzAAAAABhDzAIAAABgDDELAAAAgDHELAAAAADGELMAAAAAGEPMAgAAAGAMMQsAAACAMcQsAAAAAMYQswAAAAAYQ8wCAAAAYAwxCwAAAIAxxCwAAAAAxhCzAAAAABhDzAIAAABgDDELAAAAgDHELAAAAADGELMAAAAAGEPMAgAAAGAMMQsAAACAMcQsAAAAAMYQswAAAAAYY9sxq6quqaovV9VnF+ubq+pLVfW1qvpMVR3cvTEBAAAA4PKuzPrLSZ5Zs/7ZJB/v7qNJXk5y/04OBgAAAADrbStmVdWRJH8qyT9frCvJe5P84uLIiSR378aAAAAAAPCK7V6Z9Q+T/LUkv7tY/8Ek3+7ui4v12SQ37PBsAAAAAPAqW8asqvrTSc5195Nrtzc42pt8/lhVna6q0+cvnL/CMQEAAAAgObCNM7cnubOqPpjk9UnelNUrta6tqgOLq7OOJHlhow939/Ekx5PkD197ZMPgBQAAAADbseWVWd390919pLtvSnJPkn/b3T+e5PNJPrQ4dl+SR3dtSgAAAADI5f2a4Xp/PclPVdWZrD5D6xM7MxIAAAAAbGw7txl+V3d/IckXFq+fS3Lrzo8EAAAAABu7miuzAAAAAGBPiVkAAAAAjCFmAQAAADCGmAUAAADAGGIWAAAAAGOIWQAAAACMIWYBAAAAMIaYBQAAAMAYYhYAAAAAY4hZAAAAAIwhZgEAAAAwhpgFAAAAwBhiFgAAAABjiFkAAAAAjCFmAQAAADCGmAUAAADAGGIWAAAAAGOIWQAAAACMIWYBAAAAMIaYBQAAAMAYYhYAAAAAY4hZAAAAAIwhZgEAAAAwhpgFAAAAwBhiFgAAAABjiFkAAAAAjCFmAQAAADCGmAUAAADAGGIWAAAAAGOIWQAAAACMIWYBAAAAMIaYBQAAAMAYYhYAAAAAY4hZAAAAAIwhZgEAAAAwhpgFAAAAwBhiFgAAAABjiFkAAAAAjCFmAQAAADCGmAUAAADAGGIWAAAAAGOIWQAAAACMIWYBAAAAMIaYBQAAAMAYYhYAAAAAY4hZAAAAAIwhZgEAAAAwhpgFAAAAwBhiFgAAAABjiFkAAAAAjCFmAQAAADCGmAUAAADAGGIWAAAAAGOIWQAAAACMIWYBAAAAMIaYBQAAAMAYYhYAAAAAY4hZAAAAAIwhZgEAAAAwhpgFAAAAwBhiFgAAAABjiFkAAAAAjCFmAQAAADCGmAUAAADAGGIWAAAAAGOIWQAAAACMIWYBAAAAMIaYBQAAAMAYYhYAAAAAY4hZAAAAAIwhZgEAAAAwhpgFAAAAwBhiFgAAAABjiFkAAAAAjCFmAQAAADCGmAUAAADAGGIWAAAAAGOIWQAAAACMIWYBAAAAMIaYBQAAAMAYYhYAAAAAY4hZAAAAAIwhZgEAAAAwhpgFAAAAwBhiFgAAAABjiFkAAAAAjCFmAQAAADCGmAUAAADAGGIWAAAAAGOIWQAAAACMIWYBAAAAMIaYBQAAAMAYYhYAAAAAY4hZAAAAAIwhZgEAAAAwhpgFAAAAwBhiFgAAAABjiFkAAAAAjCFmAQAAADCGmAUAAADAGGIWAAAAAGOIWQAAAACMIWYBAAAAMIaYBQAAAMAYYhYAAAAAY4hZAAAAAIwhZgEAAAAwhpgFAAAAwBhiFgAAAABjiFkAAAAAjCFmAQAAADCGmAUAAADAGGIWAAAAAGOIWQAAAACMIWYBAAAAMMaB7Ryqqq8n+a0kv5PkYnevVNVbknwmyU1Jvp7kw9398u6MCQAAAACXd2XWj3b3u7p7ZbF+MMnj3X00yeOLNQAAAADsmqu5zfCuJCcWr08kufvqxwEAAACAzW03ZnWSX6mqJ6vq2GLvbd39YpIs/l63GwMCAAAAwCu29cysJLd39wtVdV2Sz1XVf9nuFyzi17EkedMPXnsFIwIAAADAqm1dmdXdLyz+nkvyy0luTfLNqro+SRZ/z23y2ePdvdLdK4cOHtqZqQEAAABYSlvGrKo6VFVvfOV1kj+Z5KtJTia5b3HsviSP7taQAAAAAJBs7zbDtyX55ap65fwvdPe/qaonkjxSVfcn+UaSH9u9MQEAAABgGzGru59L8iMb7H8ryft2YygAAAAA2Mh2f80QAAAAAPadmAUAAADAGGIWAAAAAGOIWQAAAACMIWYBAAAAMIaYBQAAAMAYYhYAAAAAY4hZAAAAAIwhZgEAAAAwhpgFAAAAwBhiFgAAAABjiFkAAAAAjCFmAQAAADCGmAUAAADAGGIWAAAAAGOIWQAAAABLqqruqKpnq+pMVT34Guc+VFVdVSt7Od9GxCwAAACAJVRV1yR5KMkHktyS5N6qumWDc29M8peSfGlvJ9yYmAUAAACwnG5Ncqa7n+vuC0keTnLXBuf+bpJ/kOT/7uVwmxGzAAAAAJbTDUmeX7M+u9j7rqp6d5Ibu/uzeznYaxGzAAAAAJZTbbDX332z6geSfDzJR7b8j6qOVdXpqjr90ksv7eCIlxKzAAAAAJbT2SQ3rlkfSfLCmvUbk/xwki9U1deT3Jbk5EYPge/u49290t0rhw8f3sWRxSwAAACAZfVEkqNVdXNVHUxyT5KTr7zZ3d/p7rd2903dfVOSLya5s7tP78+4q8QsAAAAgCXU3ReTPJDksSTPJHmku5+qqo9V1Z37O93mDuz3AAAAAADsj+4+leTUur2PbnL2PXsx01ZcmQUAAADAGGIWAAAAAGOIWQAAAACMIWYBAAAAMIaYBQAAAMAYYhYAAAAAY4hZAAAAAIwhZgEAAAAwhpgFAAAAwBhiFgAAAABjiFkAAAAAjCFmAQAAADCGmAUAAADAGGIWAAAAAGOIWQAAAACMIWYBAAAAMIaYBQAAAMAYYhYAAAAAY4hZAAAAAIwhZgEAAAAwhpgFAAAAwBhiFgAAAABjiFkAAAAAjCFmAQAAADCGmAUAAADAGGIWAAAAAGOIWQAAAACMIWYBAAAAMIaYBQAAAMAYYhYAAAAAY4hZAAAAAIwhZgEAAAAwhpgFAAAAwBhiFgAAAABjiFkAAAAAjCFmAQAAADCGmAUAAADAGGIWAAAAAGOIWQAAAACMIWYBAAAAMIaYBQAAAMAYYhYAAAAAY4hZAAAAAIwhZgEAAAAwhpgFAAAAwBhiFgAAAABjiFkAAAAAjCFmAQAAADCGmAUAAADAGGIWAAAAAGOIWQAAAACMIWYBAAAAMIaYBQAAAMAYYhYAAP+/vbuLtaw+7zv+ezQn05hJU2wHWw7Y8VhFJCiKsTNCTlK5qXEqSCPwha2CGgkiKm7iJk5atbSVMsJSpLiN6ubCiopsB1Sl+IXa9aiicS3iKr1oqccvrbEJ8hhimECANn5psWpC+/TiLJfpMJQzzDln8+z9+Uijvdc6a8950F9n79GXtdYBABhDzAIAAABgDDELAAAAgDHELAAAAADGELMAAAAAGEPMAgAAAGAMMQsAAACAMcQsAAAAAMYQswAAAAAYQ8wCAAAAYAwxCwAAAIAxxCwAAAAAxhCzAAAAABhDzAIAAABgDDELAAAAgDHELAAAAADGELMAAAAAGEPMAgAAAGAMMQsAAACAMcQsAAAAAMYQswAAAAAYQ8wCAAAAYAwxCwAAAIAxxCwAAAAAxhCzAAAAABhDzAIAAABgDDELAAAAgDHELAAAAADGELMAAAAAGEPMAgAAAGAMMQsAAACAMcQsAAAAAMYQswAAAAAYQ8wCAAAAYAwxCwAAAIAxxCwAAAAAxhCzAAAFlye2AAAV+ElEQVQAABhjRzGrqs6vqjur6g+r6r6q+omqellVfaqqvrI8vnSvhwUAAABgs+30zKzfSvJ73f3DSV6f5L4kNye5u7svTnL3sg0AAAAAe+Z5Y1ZVfX+SNyf5QJJ091Pd/Y0k1yS5fTns9iRv26shAQAAACDZ2ZlZr0vyRJLfqarPV9X7q+pQkld296NJsjy+Yg/nBAAAAIAdxaytJG9M8tvd/YYkT+YsLimsqpuq6nhVHX/yqSdf4JgAAAAAsLOYdTLJye6+Z9m+M9tx67GqelWSLI+Pn+nF3X1rdx/p7iOHDh7ajZkBAAAA2FDPG7O6+0+SPFxVlyy7rkjy5STHkly/7Ls+ySf2ZEIAAAAAWGzt8Li/leR3q+pgkgeS/EK2Q9hHqurGJA8lecfejAgAAAAA23YUs7r7C0mOnOFLV+zuOAAAAADw3HZyzywAAAAAeFEQswAAAAAYQ8wCAAAAYAwxCwAAAIAxxCwAAAAAxhCzAAAAABhDzAIAAABgDDELAAAAgDHELAAAAADGELMAAAAAGEPMAgAAAGAMMQsAAACAMcQsAAAAAMYQswAAAAAYQ8wCAAAAYAwxCwAAAIAxxCwAAAAAxhCzAAAAABhDzAIAAABgDDELAAAAgDHELAAAAADGELMAAAAAGEPMAgAAAGAMMQsAAACAMcQsAAAAAMYQswAAAAAYY2vVAwAAAM+45bYHz+r4ozcc3qNJAODFyZlZAAAAAIwhZgEAAAAwhpgFAAAAwBhiFgAAAABjiFkAAAAAjCFmAQAAADCGmAUAAADAGGIWAAAAAGOIWQAAAACMIWYBAAAAMIaYBQAAAMAYW6seAAAAprjltgdXPcKzvJCZjt5weA8mAYD94cwsAAAAAMYQswAAAAAYQ8wCAAAAYAwxCwAAAIAxxCwAAAAAxhCzAAAAABhDzAIAAABgDDELAAAAgDHELAAAAADGELMAAAAAGEPMAgAAAGAMMQsAAACAMbZWPQAAAExx9IbDqx7hWW657cFVjwAA+8qZWQAAAACMIWYBAAAAMIaYBQAAAMAYYhYAAAAAY4hZAAAAAIwhZgEAAAAwhpgFAAAAwBhiFgAAAABjiFkAAAAAjCFmAQAAADCGmAUAAADAGFurHgAAAHjhjt5weNUjAMC+cmYWAAAAAGOIWQAAAACMIWYBAAAAMIaYBQAAAMAYYhYAAAAAY4hZAAAAAIwhZgEAAAAwhpgFAAAAwBhiFgAAAABjiFkAAAAAjCFmAQAAADDG1qoHAACAVbnltgfP6vijNxze8++xH17IfwcAvFg4MwsAAACAMcQsAAAAAMYQswAAAAAYQ8wCAAAAYAwxCwAAAIAxxCwAAAAAxhCzAAAAABhDzAIAAABgDDELAAAAgDHELAAAAADGELMAAAAAGEPMAgAAAGCMrVUPAAAAq3L0hsNndfwttz24R5MAADvlzCwAAAAAxhCzAAAAABhDzAIAAABgDDELAAAAgDHELAAAAIANVVVXVtX9VXWiqm4+w9d/taq+XFX/parurqofWsWcpxKzAAAAADZQVR1I8r4kVyW5NMl1VXXpaYd9PsmR7v6xJHcm+Uf7O+WziVkAAAAAm+nyJCe6+4HufirJh5Jcc+oB3f3p7v72svkfk1y0zzM+i5gFAAAAsJkuTPLwKdsnl33P5cYk/2ZPJ9qBrVUPAAAAAMBK1Bn29RkPrPr5JEeS/OXn+PpNSW5Kkte85jW7Nd8ZOTMLAAAAYDOdTPLqU7YvSvLI6QdV1VuT/MMkV3f3d870F3X3rd19pLuPXHDBBXsy7HeJWQAAAACb6TNJLq6qw1V1MMm1SY6dekBVvSHJP8t2yHp8BTM+i8sMAQBgh47ecPisX3PLbQ/uwSTPeCEzAUCSdPfTVfXOJJ9MciDJB7v7S1X17iTHu/tYkn+c5PuSfLSqkuSh7r56ZUNHzAIAAADYWN19V5K7Ttv3a6c8f+u+D/U8XGYIAAAAwBhiFgAAAABjiFkAAAAAjCFmAQAAADCGmAUAAADAGGIWAAAAAGOIWQAAAACMIWYBAAAAMIaYBQAAAMAYYhYAAAAAY2ytegAAAFhnR284vOoRAGCtODMLAAAAgDHELAAAAADGeN6YVVWXVNUXTvnzrap6V1W9rKo+VVVfWR5fuh8DAwAAALC5njdmdff93X1Zd1+W5MeTfDvJx5PcnOTu7r44yd3LNgAAAADsmbO9zPCKJF/t7q8luSbJ7cv+25O8bTcHAwAAAIDTnW3MujbJHcvzV3b3o0myPL5iNwcDAAAAgNPtOGZV1cEkVyf56Nl8g6q6qaqOV9XxJ5968mznAwAAAID/62zOzLoqyee6+7Fl+7GqelWSLI+Pn+lF3X1rdx/p7iOHDh46t2kBAAAA2GhnE7OuyzOXGCbJsSTXL8+vT/KJ3RoKAAAAAM5kRzGrqs5L8jNJPnbK7t9I8jNV9ZXla7+x++MBAAAAwDO2dnJQd387yctP2/ffsv3bDQEAAABgX5ztbzMEAAAAgJURswAAAAAYQ8wCAAAAYAwxCwAAAIAxxCwAAAAAxhCzAAAAABhDzAIAAABgDDELAAAAgDHELAAAAADGELMAAAAAGEPMAgAAAGAMMQsAAACAMcQsAAAAAMYQswAAAAAYQ8wCAAAAYAwxCwAAAIAxxCwAAAAAxhCzAAAAABhDzAIAAABgDDELAAAAgDHELAAAAADGELMAAAAAGEPMAgAAAGAMMQsAAACAMcQsAAAAAMYQswAAAAAYQ8wCAAAAYAwxCwAAAIAxxCwAAAAAxhCzAAAAABhDzAIAAABgDDELAAAAgDHELAAAAADGELMAAAAAGEPMAgAAAGAMMQsAAACAMcQsAAAAAMYQswAAAAAYQ8wCAAAAYAwxCwAAAIAxxCwAAAAAxhCzAAAAABhDzAIAAABgDDELAAAAgDHELAAAAADGELMAAAAAGEPMAgAAAGAMMQsAAACAMcQsAAAAAMYQswAAAAAYQ8wCAAAAYAwxCwAAAIAxxCwAAAAAxhCzAAAAABhDzAIAAABgDDELAAAAgDHELAAAAADGELMAAAAAGEPMAgAAAGAMMQsAAACAMcQsAAAAAMYQswAAAAAYQ8wCAAAAYAwxCwAAAIAxxCwAAAAAxhCzAAAAABhDzAIAAABgDDELAAAAgDHELAAAAADGELMAAAAAGEPMAgAAAGAMMQsAAACAMcQsAAAAAMYQswAAAAAYQ8wCAAAAYAwxCwAAAIAxxCwAAAAAxhCzAAAAABhDzAIAAABgDDELAAAAgDHELAAAAADGELMAAAAAGEPMAgAAAGAMMQsAAACAMcQsAAAAAMYQswAAAAAYQ8wCAAAAYAwxCwAAAIAxxCwAAAAAxhCzAAAAABhDzAIAAABgDDELAAAAgDHELAAAAADGELMAAAAAGEPMAgAAAGAMMQsAAACAMcQsAAAAAMYQswAAAAAYQ8wCAAAAYAwxCwAAAIAxxCwAAAAAxhCzAAAAABhDzAIAAABgDDELAAAAgDHELAAAAADGELMAAAAAGEPMAgAAAGAMMQsAAACAMcQsAAAAAMYQswAAAAAYQ8wCAAAAYAwxCwAAAIAxxCwAAAAAxhCzAAAAABhjRzGrqn6lqr5UVfdW1R1V9b1Vdbiq7qmqr1TVh6vq4F4PCwAAAMBme96YVVUXJvmlJEe6+0eTHEhybZL3JHlvd1+c5OtJbtzLQQEAAABgp5cZbiV5SVVtJTkvyaNJ3pLkzuXrtyd52+6PBwAAAADPeN6Y1d1/nOQ3kzyU7Yj1zSSfTfKN7n56Oexkkgv3akgAAAAASHZ2meFLk1yT5HCSH0xyKMlVZzi0n+P1N1XV8ao6/uRTT57LrAAAAABsuJ1cZvjWJA929xPd/WdJPpbkJ5Ocv1x2mCQXJXnkTC/u7lu7+0h3Hzl08NCuDA0AAADAZtpJzHooyZuq6ryqqiRXJPlykk8neftyzPVJPrE3IwIAAADAtp3cM+uebN/o/XNJvri85tYkfy/Jr1bViSQvT/KBPZwTAAAAALL1/Ick3X00ydHTdj+Q5PJdnwgAAAAAnsNOLjMEAAAAgBcFMQsAAACAMcQsAAAAAMYQswAAAAAYQ8wCAAAAYAwxCwAAAIAxxCwAAAAAxhCzAAAAABhDzAIAAABgDDELAAAAgDHELAAAAADGELMAAAAAGEPMAgAAAGAMMQsAAACAMcQsAAAAAMYQswAAAAAYQ8wCAAAAYAwxCwAAAIAxxCwAAAAAxhCzAAAAABhDzAIAAABgDDELAAAAgDHELAAAAADGELMAAAAAGEPMAgAAAGAMMQsAAACAMcQsAAAAAMYQswAAAAAYQ8wCAAAAYAwxCwAAAIAxxCwAAAAAxhCzAAAAABhDzAIAAABgDDELAAAAgDHELAAAAADGELMAAAAAGEPMAgAAAGAMMQsAAACAMcQsAAAAAMYQswAAAAAYQ8wCAAAAYAwxCwAAAIAxxCwAAAAAxhCzAAAAABhDzAIAAABgDDELAAAAgDHELAAAAADGELMAAAAAGEPMAgAAANhQVXVlVd1fVSeq6uYzfP3PVdWHl6/fU1Wv3f8p/19iFgAAAMAGqqoDSd6X5Koklya5rqouPe2wG5N8vbv/YpL3JnnP/k75bGIWAAAAwGa6PMmJ7n6gu59K8qEk15x2zDVJbl+e35nkiqqqfZzxWcQsAAAAgM10YZKHT9k+uew74zHd/XSSbyZ5+b5M9xy2VvnNAQAAAFiZM51h1S/gmFTVTUluWja/U1X3voB5LtnJQWIWAAAAwGY6meTVp2xflOSR5zjmZFVtJfkLSf709L+ou29NcmuSVNXx7j5ytsNU1fGdHOcyQwAAAIDN9JkkF1fV4ao6mOTaJMdOO+ZYkuuX529P8vvd/awzs/aTM7MAAAAANlB3P11V70zyySQHknywu79UVe9Ocry7jyX5QJJ/XlUnsn1G1rWrm3ibmAUAAACwobr7riR3nbbv1055/j+TvOMs/9pbX+A4O3qdywwBAAAA2DXL/bP27HViFgAAAABjiFkAAAAAnLOqurKq7q+qE1V181m87oNV9XhV3buT48UsAAAAAM5JVR1I8r4kVyW5NMl1VXXpDl9+W5Ird/q9xCwAAAAAztXlSU509wPd/VSSDyW5Zicv7O4/yPZvStwRMQsAAACAc3VhkodP2T657Nt1YhYAAAAA56rOsK/34huJWQAAAACcq5NJXn3K9kVJHtmLbyRmAQAAAHCuPpPk4qo6XFUHk1yb5NhefCMxCwAAAIBz0t1PJ3lnkk8muS/JR7r7Szt5bVXdkeQ/JLmkqk5W1Y3/v+O3znVYAAAAAOjuu5Lc9QJed93ZHO/MLAAAAADGELMAAAAAGEPMAgAAAGAMMQsAAACAMcQsAAAAAMYQswAAAAAYQ8wCAAAAYAwxCwAAAIAxxCwAAAAAxhCzAAAAABhDzAIAAABgDDELAAAAgDHELAAAAADGELMAAAAAGEPMAgAAAGAMMQsAAACAMcQsAAAAAMYQswAAAAAYQ8wCAAAAYAwxCwAAAIAxqrv375tVPZHka2f40g8k+a/7NgirZr03h7XeLNZ7c1jrzWK9N4e13izWe3NY6/33Q919waqHWGf7GrOec4iq4919ZNVzsD+s9+aw1pvFem8Oa71ZrPfmsNabxXpvDmvNOnKZIQAAAABjiFkAAAAAjPFiiVm3rnoA9pX13hzWerNY781hrTeL9d4c1nqzWO/NYa1ZOy+Ke2YBAAAAwE68WM7MAgAAAIDntfKYVVVXVtX9VXWiqm5e9Tzsnqr6YFU9XlX3nrLvZVX1qar6yvL40lXOyO6pqldX1aer6r6q+lJV/fKy35qvmar63qr6T1X1n5e1vmXZf7iq7lnW+sNVdXDVs7I7qupAVX2+qv71sm2t11RV/VFVfbGqvlBVx5d93sfXVFWdX1V3VtUfLp/fP2G9109VXbL8TH/3z7eq6l3Wen1V1a8s/0a7t6ruWP7t5rObtbLSmFVVB5K8L8lVSS5Ncl1VXbrKmdhVtyW58rR9Nye5u7svTnL3ss16eDrJ3+7uH0nypiS/uPw8W/P1850kb+nu1ye5LMmVVfWmJO9J8t5lrb+e5MYVzsju+uUk952yba3X21/p7stO+TXu3sfX128l+b3u/uEkr8/2z7n1XjPdff/yM31Zkh9P8u0kH4+1XktVdWGSX0pypLt/NMmBJNfGZzdrZtVnZl2e5ER3P9DdTyX5UJJrVjwTu6S7/yDJn562+5okty/Pb0/ytn0dij3T3Y929+eW5/892/8gvjDWfO30tv+xbH7P8qeTvCXJnct+a70mquqiJH8tyfuX7Yq13jTex9dQVX1/kjcn+UCSdPdT3f2NWO91d0WSr3b312Kt19lWkpdU1VaS85I8Gp/drJlVx6wLkzx8yvbJZR/r65Xd/WiyHT+SvGLF87AHquq1Sd6Q5J5Y87W0XHb2hSSPJ/lUkq8m+UZ3P70c4v18ffzTJH83yf9etl8ea73OOsm/rarPVtVNyz7v4+vpdUmeSPI7y2XE76+qQ7He6+7aJHcsz631GuruP07ym0keynbE+maSz8ZnN2tm1TGrzrDPr1eEwarq+5L8yyTv6u5vrXoe9kZ3/6/lcoWLsn2W7Y+c6bD9nYrdVlU/l+Tx7v7sqbvPcKi1Xh8/1d1vzPYtIH6xqt686oHYM1tJ3pjkt7v7DUmejMvM1tpyj6Srk3x01bOwd5Z7n12T5HCSH0xyKNvv6afz2c1oq45ZJ5O8+pTti5I8sqJZ2B+PVdWrkmR5fHzF87CLqup7sh2yfre7P7bstuZrbLkk5d9l+z5p5y+nsyfez9fFTyW5uqr+KNu3AnhLts/UstZrqrsfWR4fz/Y9dS6P9/F1dTLJye6+Z9m+M9txy3qvr6uSfK67H1u2rfV6emuSB7v7ie7+syQfS/KT8dnNmll1zPpMkouX36xwMNunvR5b8UzsrWNJrl+eX5/kEyuchV203EfnA0nu6+5/csqXrPmaqaoLqur85flLsv2PpvuSfDrJ25fDrPUa6O6/390Xdfdrs/0Z/fvd/TdirddSVR2qqj//3edJ/mqSe+N9fC11958kebiqLll2XZHky7He6+y6PHOJYWKt19VDSd5UVect/z7/7s+2z27WSnWv9uzCqvrZbP9f3gNJPtjdv77Sgdg1VXVHkp9O8gNJHktyNMm/SvKRJK/J9hvtO7r79JvEM1BV/aUk/z7JF/PMvXX+Qbbvm2XN10hV/Vi2bxx6INv/U+Qj3f3uqnpdts/eeVmSzyf5+e7+zuomZTdV1U8n+Tvd/XPWej0t6/rxZXMryb/o7l+vqpfH+/haqqrLsv3LHQ4meSDJL2R5X4/1XitVdV6271X8uu7+5rLPz/aaqqpbkvz1bP+28c8n+ZvZvkeWz27WxspjFgAAAADs1KovMwQAAACAHROzAAAAABhDzAIAAABgDDELAAAAgDHELAAAAADGELMAAAAAGEPMAgAAAGAMMQsAAACAMf4PCklqSxYfgfsAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 1440x1440 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#plt.imshow(img_data[:,:,[2,1,0]])\n",
    "plt.imshow(tmp[:,:], alpha=.7)#alpha -> 이미지 얼마나 선명한가. heatmap_avg[:,:,0] -> 1:r_knee\n",
    "fig = matplotlib.pyplot.gcf()\n",
    "cax = matplotlib.pyplot.gca()\n",
    "fig.set_size_inches(20, 20)\n",
    "fig.subplots_adjust(right=0.93)\n",
    "cbar_ax = fig.add_axes([0.95, 0.15, 0.01, 0.7])\n",
    "#_ = fig.colorbar(oriImg, cax=cbar_ax)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "plt.rcParams['figure.figsize'] = (20,20)\n",
    "plt.rcParams['image.interpolation'] = 'nearest'\n",
    "plt.rcParams['image.cmap'] = 'gray'\n",
    "\n",
    "edges = [[0,1],[1,2],[2,3],[3,4],[0,5],[5,6],[6,7],[7,8],[0,9],[9,10],[10,11],[11,12],[0,13],[13,14],[14,15],[15,16],[0,17],[17,18],[18,19],[19,20]]\n",
    "\n",
    "outpath = './output_viz/'\n",
    "if not os.path.isdir(outpath):\n",
    "    os.makedirs(outpath)\n",
    "\n",
    "# Input data paths\n",
    "paths = ['synth1/', 'synth2/', 'synth3', 'synth4']\n",
    "paths = ['manual_test/', 'manual_train/']\n",
    "inpath = paths[0]\n",
    "\n",
    "files = sorted([f for f in os.listdir(inpath) if f.endswith('.json')])\n",
    "for f in files[-2:-1]:\n",
    "    with open(inpath+f, 'r') as fid:\n",
    "        dat = json.load(fid)\n",
    "\n",
    "    # Each file contains 1 hand annotation, with 21 points in\n",
    "    # 'hand_pts' of size 21x3, following this scheme:\n",
    "    # https://github.com/CMU-Perceptual-Computing-Lab/openpose/blob/master/doc/output.md#hand-output-format\n",
    "    # The 3rd column is 1 if valid:\n",
    "    pts = np.array(dat['hand_pts'])\n",
    "    invalid = pts[:,2]!=1\n",
    "    \n",
    "    # Left hands are marked, but otherwise follow the same point order\n",
    "    is_left = dat['is_left']\n",
    "\n",
    "    dat['hand_pts'] = pts.tolist()\n",
    "\n",
    "    # Plot annotations\n",
    "    plt.clf()\n",
    "    im = plt.imread(inpath+f[0:-5]+'.jpg')\n",
    "    plt.imshow(im)\n",
    "    for p in range(pts.shape[0]):\n",
    "        if pts[p,2]!=0:\n",
    "            plt.plot(pts[p,0], pts[p,1],'r.')\n",
    "            plt.text(pts[p,0], pts[p,1], '{0}'.format(p))\n",
    "    for ie, e in enumerate(edges):\n",
    "        if np.all(pts[e,2]!=0):\n",
    "            rgb = matplotlib.colors.hsv_to_rgb([ie/float(len(edges)),1.0,1.0])\n",
    "            plt.plot(pts[e,0],pts[e,1],color=rgb)\n",
    "    if is_left:\n",
    "        plt.text(10, 30, 'left', color='r', fontsize=24)\n",
    "\n",
    "    # These fields are only present in manual annotations.\n",
    "    # head_box (or head_size) is used to normalize size wrt. MPII head size for testing\n",
    "    if 'head_box' in dat:\n",
    "        head_box = np.array(dat['head_box'])\n",
    "        head_size = dat['head_size']\n",
    "        bpts = np.array(dat['mpii_body_pts'])\n",
    "\n",
    "        for p in range(bpts.shape[0]):\n",
    "            plt.plot(bpts[p,0], bpts[p,1],'g.',ms=10)\n",
    "            plt.text(bpts[p,0], bpts[p,1], '{0}'.format(p), color='w')\n",
    "        plt.plot(head_box[0,0], head_box[0,1],'b*',ms=10)\n",
    "        plt.plot(head_box[1,0], head_box[1,1],'b*',ms=10)\n",
    "        plt.plot(head_box[0:2,0], head_box[[0,0],1],'b')\n",
    "        plt.plot(head_box[0:2,0], head_box[[1,1],1],'b')\n",
    "        plt.plot(head_box[[0,0],0], head_box[0:2,1],'b')\n",
    "        plt.plot(head_box[[1,1],0], head_box[0:2,1],'b')\n",
    "\n",
    "    plt.axis('off')\n",
    "    plt.savefig(outpath+f[0:-5]+'.jpg', bbox_inches='tight')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}