{"id":36993,"date":"2025-12-10T14:27:29","date_gmt":"2025-12-10T13:27:29","guid":{"rendered":"https:\/\/www.graviton.at\/letterswaplibrary\/large-scale-image-dataset-of-perceptual-hashing\/"},"modified":"2025-12-10T14:27:29","modified_gmt":"2025-12-10T13:27:29","slug":"large-scale-image-dataset-of-perceptual-hashing","status":"publish","type":"post","link":"https:\/\/www.graviton.at\/letterswaplibrary\/large-scale-image-dataset-of-perceptual-hashing\/","title":{"rendered":"Large-scale Image Dataset Of Perceptual Hashing?"},"content":{"rendered":"

<\/p>\n

\n

‘Our dataset contains 1 200 original images’ which is not that many<\/p>\n

Do you know of a big dataset of
URL, date first, date last, phash (or other well used perceptual hash)<\/p>\n

for millions\/billions of images<\/p>\n

It seems to be the sort of thing that would be <\/p>\n

    \n
  1. \n

    useful. ‘this photo first posted here’ is a useful thing to know.<\/p>\n<\/li>\n

  2. \n

    Fairly small. Those above would be about a kb per image. a billion of those is a terabyte. <\/p>\n<\/li>\n

  3. \n

    A complete pain to make the first time. <\/p>\n<\/li>\n<\/ol>\n

    It would not get you images of the same scene or massively modified but the tiny size of the data means thats a trade off. <\/p>\n<\/div>\n

    submitted by \/u\/cavedave <\/a>
    [link]<\/a><\/span> [comments]<\/a><\/span><\/p>