A script deletes similar images through vectorization and clustering of images.
The following options are avalidable for vectorization:
- Middle layer of Xception by keras-application
- wavelet hash by Imagehash
python >= 3.6.4
pip install -r requairements.txt
python main.py INPUT_DIR OUTPUT_DIR NUM_SAMPLE
[-h] [--extractor {Xception,whash}]