This app detects objects in photos using Computer Vision and automatically adds keywords to JPEG Exif metadata. The keywords are sorted by the area of objects. It can tag a single photo, or recursively tag photos in a directory. The app is written in Python and is based on an open-source Computer Vision library cvlib from arunponnusamy. The library uses a pre-trained YOLO v3 model, which is capable of detecting 80 different objects.
pip install opencv-python tensorflow cvlib piexif
Detect objects and add keywords to example.jpg.
pyphototagger.py -f example.jpg
Detect objects and add keywords to example.jpg with verbose output.
pyphototagger.py -v -f example.jpg
Detect objects and add keywords to all *.jpg and *.jpeg files recursively found in example/photos directory.
pyphototagger.py -d example/photos
Print help
pyphototagger.py -h