demo

Quik Start

1.Check all the dependencies installed

pip install -r requirements.txt

2.Download YOLOv3 parameters

cd detector/YOLOv3/weight/
wget https://pjreddie.com/media/files/yolov3.weights
wget https://pjreddie.com/media/files/yolov3-tiny.weights
cd ../../../

3.Download deepsort parameters ckpt.t7

cd deep_sort/deep/checkpoint
# download ckpt.t7 from
https://drive.google.com/drive/folders/1xhG0kRH1EX5B9_Iz8gQJb7UNnn_riXi6 to this folder
cd ../../../

4.Compile nms module

cd detector/YOLOv3/nms
sh build.sh
cd ../../..

5.Create Face recognition database(db.npz)

pip install face_recognition
usage: face_database.py(dont't forget to change face database path in the python file)

6.Run demo

(for RGB-D video)
usage: python yolov3_deepsort.py RGB_VIDEO_PATH DEPTH_VIDEO_PATH FACE_DATA_PATH
                            [--help]
                            [--frame_interval FRAME_INTERVAL]
                            [--config_detection CONFIG_DETECTION]
                            [--config_deepsort CONFIG_DEEPSORT]
                            [--display]
                            [--display_width DISPLAY_WIDTH]
                            [--display_height DISPLAY_HEIGHT]
                            [--save_path SAVE_PATH]          
                            [--cpu]   
(for RGB video)                         
usage: python yolov3_deepsort_rgb.py RGB_VIDEO_PATH FACE_DATA_PATH
                            [--help]
                            [--frame_interval FRAME_INTERVAL]
                            [--config_detection CONFIG_DETECTION]
                            [--config_deepsort CONFIG_DEEPSORT]
                            [--display]
                            [--display_width DISPLAY_WIDTH]
                            [--display_height DISPLAY_HEIGHT]
                            [--save_path SAVE_PATH]          
                            [--cpu]   

Test Demo

You will find more details in industry_test.ipynb