Improve image duplicate processing

  • 1
  • Idea
  • Updated 1 month ago
  • (Edited)
As an amateur photographer, I've been using Lightroom and Photoshop for
years.  After all the bug fixes and new features introduced over this time, one of my frustrations is the lame image import function.  When is Adobe going to provide the capability to look at the actual images (i.e., pixels, phash, etc.) to determine if an imported image is a duplicate or not?  The import function actually looks for metadata text differences - not the images.  I don't believe this comparison method has ever changed.  Furthermore, why isn't automated identification of duplicate images in the catalog already included in the base software?  And while I'm at it, what about an additional search criteria for monochrome vs.color images?
Photo of Richard Tatum

Richard Tatum

  • 4 Posts
  • 0 Reply Likes
  • frustrated

Posted 1 month ago

  • 1
Photo of Brian Pierce

Brian Pierce

  • 73 Posts
  • 24 Reply Likes
If you''re importing duplicate images accidentally then It would suggest that your workflow needs attention.  If the import function were to work on the actual image pixels then thats going to get very processor intensive and slow importing - it also may be counter productive. many people take multiple shots of the same /similar image - what level on match would be considered a duplicate .
Photo of Richard Tatum

Richard Tatum

  • 4 Posts
  • 0 Reply Likes
I am not thinking only of duplicates in the batch I am importing, but duplicates in the batch, as well as my entire catalog.  This is a function that could run in the background - after an import is done.  And while its running, it could also look for duplicates that might already exist in the catalog.  I understand that it may be slow, but so what?  This is what computers do best.  My computer doesn't need me waiting around while it is processing the data.  Let it run overnight!   I would expect this function to have various "switches" that, for example, could ignore duplicates being imported and the algorithm would be far more sophisticated than just comparing pixels (e.g., look up "phash"). Also, it could look for similar images, not just those that perfectly match. The degree of similarity should be adjustable by the user.
I scan a lot of old negatives, also.  Many of these I scanned years ago and are already in the catalog.  During import scans usually have very little metadata, so finding a duplicate using the current text-based method alone doesn't work well for me and, more importantly, it doesn't work at all if the image is already in the catalog.


Photo of Yves Crausaz

Yves Crausaz

  • 753 Posts
  • 106 Reply Likes
Avant de vouloir de nouveaux outils, apprenez à bien utiliser ceux qui sont déjà présents dans LR !
Pour les doublons à l'importation, cochez la case ne pas importer les images déjà importées.
Pour les multiples images prises d'un même sujet , en rafale, par exemple, utiliser la fonction d’empilement automatique etc....
Ensuite, usez des tags rejetée, conservée etc....
Enfin, utilisez aussi les critères de tri personnalisés et les fonctions du suppressions assistées etc....
Classer vos images dans des collections, une collection par projet....
Vous verrez, il y a déjà de quoi faire !

(Edited)
Photo of Richard Tatum

Richard Tatum

  • 4 Posts
  • 0 Reply Likes
I understand how Lightroom import works.  I suppose what I am saying is that the duplicate finding function in Lightroom is very limited.  Why isn't it checking the entire catalog for duplicates, not just those being imported.  I also believe that machines should be doing the work - not me!  If I were a professional photographer I would be far more careful about labeling my images.  I don't want to spend all my time tagging and rating images.  Remember that Lightroom is supposed to find *duplicate images*.  Instead it is finding images that contain the same metadata.  This is not the same.

Photo of Tom Mickow

Tom Mickow

  • 514 Posts
  • 210 Reply Likes
Afraid nothing beats good organization & workflow.  I have 115,000+ images in my catalog and can guarantee with 100% certainty that there isn't a single duplicate in there.

I have no experience with them (because my process ensures that I don't have any duplicates! :) ), but there are a couple Lr plug-ins to help identify duplicates.  You should be able to find them with a quick Google search, if you're interested.  Pretty sure they're all metadata based though.  Can't imagine anyone is doing any other type of comparison.
Photo of Richard Tatum

Richard Tatum

  • 4 Posts
  • 0 Reply Likes
Yeah, I could probably update my workflow, but there isn't any reason why a lot of this stuff can't be done by software.  Adobe hasn't significantly improved most of their products in a long time.  I don't follow this closely, but adding back facial recognition was a really useful change.  There are applications that look for similarities and duplicates in data.  That's how a teacher can use internet tools to find out that a student has plagiarized a paper.  It's just math!