Detect duplicates by EXIF tag value

[aNNiMON]

Feature Description

It would be nice to have the ability to detect duplicate images using EXIF tags, such as:

• Date taken
• GPS, as it unlikely identical in two separately taken photos
• Date digitized

There is already an open issue #890, but for me, the file creation/modification date can easily be changed with simple actions, while EXIF metadata is usually preserved. This could be useful when the original photos are named something like IMG_20260215_* or DSC03228* and are then renamed according to a different pattern, such as YYYYMMDD_*. On top of that, the photos are modified via color correction and/or cropping. This causes the Similar Images tool to fail to detect duplicates, and the existing Size, Hash, and Name checks aren't helpful in such cases. However, matching by EXIF tag would solve the issue.

[Krzysiu]

Checking by date is as one of the first filtering when dealing with clutter in photos - together with binary duplicate checking, then image binary duplicate (comparing just contents, ignoring exif and other blocks). And it's a fine method, but one must remember that basically every camera since 15-20 years can take more than one photo per second, so for such tasks, it's always a good practice to check subsec (if they exist)

And I would advise against using GPS. It's identical is more than likely, it's rather sure because GPS position is usually cached, so with some cameras you can get GPS tag of place you left 15 minutes ago (because it couldn't get a new fix and used the last one).

At last, but not least, in case of most cameras taken IS digitized. The first is date when the photo was taken and digital cameras digitize it the same moment, as it was taken. I look at my photos from varierty of phones, click-and-shoot cameras, bridge cameras and DSLRs and it's always the same. Maybe some cameras differentiate it, e.g. by using digitized for the finish of saving operation.

Overall that would be wonderful, but it would need a lot of work - preferably something to display tags, something to fetch them, logic of comparing them etc. If we'd ever get that - great. For now, I refined for public use my old tool: 💡 https://github.com/Krzysiu/exifDupeFinder/ 💡 (take note 1) it's command line tool, 2) not an alternative to Czkawka - it has totally different scope) - it does exactly what you need. Compares files by metadata tags chosen by you. It can also compare imagehash - the one that I mentioned earlier. I hope this will get integrated into czkawka as well, but logic of getting just pixel data from image needs to support all types of various exotic file formats, so being realistic - I'm not counting on it.

[aNNiMON]

@Krzysiu awesome! Using this script I was able to find almost 30 duplicate photos, that

• were cropped and saved under new names
• a series of photos that were combined into a single panorama photo
  None of them weren't previously detected by czkawka.

For me, it's better to have false-positives in this case, as long as they're manually reviewed.
Thank you!