Image Edition

Finding image duplicates in collections

Digikam

• The best tool by far is Digikam, it is very convenient to quickly find duplicates and erase them. You should usually select a similarity factor of 65-70%.
• Finding duplicates only work if fingerprints have been generated first. This can be done using Tools -> Maintenance -> Fingerprints.
• The main default of Digikam is that it cannot work on folders that are not part of its "collections" (although it is quite easy to add a folder to a collection). Also it seems to create a hidden (starting with a dot) folder in some directories, to implement trashes.
• Be careful that an album is not the same as a collection. A collection (or root folder) has several albums and is configured via Settings -> Configure Digikam.

Geeqie

• A tool that can work as well is Geeqie (in Portage). However I found out its behavior to be a bit strange and thus Digikam is better (also, it's not possible to view easily found duplicates to confirm them, which is very problematic, and Digikam does that very well).
• Choose the folder where you want to search for duplicates (you can select a top-level folder and scan recursively), then choose Compare by Similarity (or Similarity - High).
• Once it gives you a list of found duplicates, you can select them all, choose "View in new window", then click on the new window to slideshow the images.
• The most convenient way to remove the duplicates is for me to move all found images to a new place (other folder) then open the folder in Gwenview and remove duplicates manually.
• Digikam also has some duplicate searching features (not tried yet).