imageAnalyzer

Analyzes quality of images in pdf's

requirements

• python (from Python 3.7)

• matlab (from R2019b)

• poppler pdfimages (0.68.0)

• exif==1.3.5

• haishoku==1.1.8

• layoutparser[effdet]==0.3.4

• numpy==1.21.6

• opencv_python==4.6.0.66

• Pillow==9.3.0

• protobuf==4.21.12

• pyenchant==3.2.2

• PyMuPDF==1.20.2

• pytesseract==0.3.10

• reportlab==3.6.11

• wcag_contrast_ratio==0.9

• importlib-metadata==5.2.0

install (Windows 10)

Tested Windows 10, Python 3.10 and Matlab R2022b

dictionary

files 'de_DE_frami.aff' and 'de_DE_frami.dic' source need to be moved to pyenchant folder [...]\enchant\data\mingw64\share\enchant\hunspell

run

To analyse PDF run main.py file with path of pdf file as parameter

example: python main.py myTestPdf.pdf

install (Docker)

Tested with Linux, Python 3.7 and Matlab R2019b

build docker image

From the repository root, build the docker image:

docker build -t schall/image_analyzer:git -f docker/Dockerfile .

run

Run it while mounting the host's MATLAB installation into the container (must not be read-only because we have to build the MATLAB python engine):

docker run -it --rm -v /usr/local/MATLAB/:/usr/local/MATLAB -v /path/to/input_dir:/input:ro -v /path/to/output_dir:/output schall/image_analyzer:git

Now navigate to main/src, run the main.py with path of pdf file as parameter , and finally export the output:

cd main/src
python3 main.py /input/some.pdf
cp ../../output/bildanalyse_report.pdf /output/some_report.pdf