Hacker News Reader

I should think you could do this with fingerprinting the photo similar to how music is fingerprinted for things like Shazam or MusicBrainz. I used to work for MusicIP, which I believe developed the fingerprinting system MusicBrainz is using.

View on HN

Thanks for the suggestion - I looked into music recognition algorithms early on but struggled to adapt them for image use. But I'll revisit them.

Are there "fuzzy" fingerprint algorithms that tolerate some variation in pixel color / hues, edges, imaging quality, etc?

There are image-representation versions of wavelets that would work well in that context, with some tolerance/quantization of the frequency representation to accommodate fuzzy edges, and likewise for nearby hues.

Perceptual color representation gets a bit harder but if you're only looking at gamut differences on cameras/screens/printed media I think it's feasible.

Alternatively, if you know a lot about the source image you can train a NN for the specific application.

Reply on news.ycombinator.com