How can I find duplicate photos when some recovered JPEGs are partially corrupted?

Asked 6/14/2020

3 views

2 answers

0

I’m consolidating a large family photo archive recovered from damaged hard drives, CDs, and floppy disks. I ended up with many duplicate files, including partially corrupted JPEGs. Standard duplicate-photo tools work well for intact images, but they often fail on damaged files because the recovered image may only contain part of the original photo.

I’d like a way to identify whether a corrupted JPEG is a duplicate of a good copy I already have, ideally by comparing image data in a more direct way than typical low-resolution or content-aware matching. Is there software or a practical method for this?

Originally by Photography Stack Exchange contributor. Source · Licensed CC BY-SA 4.0

Photography Stack Exchange contributor

6y ago

2 Answers

1

The free software AllDup allows to check duplicates only using a small sector in the top left of the image thus partly solving the problem

Originally by user92591. Source · Licensed CC BY-SA 4.0

user92591

4y ago

0

AI Answer

Generated from our catalog & community — verify before relying on it.

For exact duplicate files, the most reliable approach is file hashing rather than visual comparison. Generate a checksum such as SHA for each file and compare the hashes; matching hashes indicate the files are effectively identical with very high probability.

For partially corrupted JPEGs, exact hashing usually won’t help because the damaged file is no longer byte-for-byte the same. In that case, a duplicate finder that can compare only part of the image may work better. Community suggestions include:

  • AllDup: can compare duplicates using just a small sector in the top-left of the image, which may help when only part of a recovered JPEG is intact.
  • ACDSee Pro: can find similar images and lets you adjust the allowed degree of difference.

A true pixel-by-pixel match is generally less practical for corrupted JPEGs because corruption changes the decoded image structure, not just a few pixels. So the best workflow is usually:

  1. use SHA hashes to remove exact duplicate files,
  2. then use a similarity tool such as AllDup or ACDSee Pro for partial/corrupted image matches.

UniqueBot

AI

6y ago

Your Answer