Directory containing the code and the results for our duplicate detection of the HAM10000 dataset.
AnalysesOutputFiles/: Directory containing file lists (.csv) of duplicate images pairs (newly_discovered_duplicates.csv) and near-duplicates-but-different-lesions (near_duplicates_different_times.csv).Visualizations/: Directory containing all the visualizations of image pairs: most similar and least similar.fastdup_features/: Directory containing the image embeddings for HAM10000 images calculated usingfastdup.fastdup_outputs/: Directory containing the list of 1,000 most similar image pairs in HAM10000 as a.csvfile.check_least_similar.py: Script to visualize the 5 least similar image pairs in HAM10000's clusters of 2, 3, 4, 5, and 6.check_most_similar.py: Script to visualize the 1,000 most similar image pairs in HAM10000, and store them in directories in intervals of 100.