Category talk:Images using filenames with Facebook photo identities

Log of files added with predicted Facebook links to original[edit]

The Facebook links below will show the original photo so long as it is set to public view and has not been deleted since being uploaded. The log includes deleted files to provide an easy audit trail.

Note, this list is updated manually from a text log. If you are looking for a file which was categorized recently (within the last 24 hours), then it may take a while longer to appear here, however for files categorized on 18 November onwards, the Facebook link will be in the edit comment history. --Fæ (talk) 17:48, 18 November 2013 (UTC)[reply]

Facebook or Flickr?[edit]

These filenames with many digits separated by spaces and then a letter, sometimes with an additional parenthesized number, always scream Flickr at me. That must be because I dont have much experience with Facebook imports, but how are they distinguishable? The following File:423093 10150619590174143 1635341394 n (7874712114).jpg caught my eye, and it is indeed a valid import from Flickr, categorized and everything. Is this category contaminated with false positives? -- Tuválkin ✉ 11:28, 18 November 2013 (UTC)[reply]

Flickr use different name structures but one of them is a unique photo-ID, this file was moved from Facebook to Flickr and is a bit questionable: According to Exif data it took two minutes to upload it to facebook, download it from facebook just to upload it to Flickr. Recent uploads of this user are from Facebook, those files do not contain usable Exif data. Impossible to verify these as genuine photos made by the Flickr user. --Denniss (talk) 11:46, 18 November 2013 (UTC)[reply]

It's one of those areas of where to judge good faith. The image is on a suitable licence on Flickr and displayed publicly with no other evidence of alternative claims of copyright via a Google Image or Tineye search. The only reason we suspected this image was a poor file name, if someone were deliberately Flickrwashing, this would be a bit of a give away.

With regard to false positives, in a sample of the first 200, none is a false positive so far. The matching is very simple, just checking for any file name including the pattern <6 digits>_<17 digits>_<6 digits>, which appears to be a minimum default for Facebook IDs (fids). If someone can produce counter-examples I would be happy to reconsider what to search for. In general I would rather a conservative approach missed possible candidates rather than include false positives, after all we can refine this matching later on. A possible future refinement would be to understand the fids structure, I suspect we can work out a pattern such that we can tell when sets of images are coming from the same Facebook account, consequently finding a pattern of good or bad images can then influence a decision for the rest of that set. --Fæ (talk) 12:21, 18 November 2013 (UTC)[reply]

Apologies for the false alarm. Flickr filenames are also a bit of a salad, but very different in all details, except the underscore+letter end. -- Tuválkin ✉ 15:19, 18 November 2013 (UTC)[reply]

As mentioned, a bit of thought about the fids structure has lead to Facebook links being added to the simple list above. --Fæ (talk) 16:39, 18 November 2013 (UTC)[reply]

New identification methods for Facebook images[edit]

https://twitter.com/oasace/status/1149181539000864769 "I noticed a structural abnormality when looking at a hex dump of an image file from an unknown origin only to discover it contained what I now understand is an IPTC special instruction. Shocking level of tracking.." Not clear whether the tracking info actually relates to the downloading user. Nemo 14:14, 13 July 2019 (UTC)[reply]

Category talk:Images using filenames with Facebook photo identities

Log of files added with predicted Facebook links to original[edit]

Facebook or Flickr?[edit]

New identification methods for Facebook images[edit]

Navigation menu

Search