Dmytro Galko · open-source tools
The same photo twice, found by what it looks like
Byte-comparison finds only files that are literally identical, and misses the case everyone actually has: the same photo at 4K and at 1920, the JPEG re-saved by a phone, the export that changed nothing but the metadata. This groups by what a picture looks like, tells you which copy to keep, and never deletes anything.
31,662 JPEGs from a real photo folder hashed in 37 seconds on 16 threads. 23 tests, plus a benchmark that rebuilds its own test folder so the grouping can be checked rather than believed.
find duplicate photos perceptual hash, near duplicate image finder cli, resized recompressed same picture, rust image dedupe 31000 photos.
If that is your week and you would rather someone else did it, write to hello@dkautomation.dev or open an issue.