Duplicated image finder

TheSaint · 2012-07-29 15:54:58

Helllo,
I use to back a lot of photo, and most of them are saved in different places. I'm trying to sort out these things and recover a bunch of Gigabytes.
Unfortunately I couldn't find a specific software that may look recursively in a directory (or a list of directories) and show me what could be duplicated.
Something is done by Geeqie, but it isn't specific for that purpose and it takes some stunt to perform a comparison.
I hope the developers will take a review and allow some plugin, in order to expand the capabilities into that way.
Another way, I should find out how the Geeqie collections are made of, and try some alchemy with python to generate them, recursively.
Actually, I install wine to perform this task. I just wonder how a freeware program has taken libraries from GTK and compile into a windows executable .

karol · 2012-07-29 16:25:51

Images are just files, try e.g. rmlint.

TheSaint · 2012-07-29 16:35:21

karol wrote:

Images are just files, try e.g. rmlint.

Some images were resized and resolution reduced, for mailing purpose. I'd like to make a choise on which I'll keep.
So for those images that are identical, I did the file comparison, using fdupes.
But like most of image comparison they use to scale down the image into a thumbnail size and compare them. Another procedure is to collect all of then and compare one by one against all the rest.

karol · 2012-07-29 16:38:40

Did you change their filenames too?

TheSaint · 2012-07-30 10:12:45

Some was changed because the transfer with gwenview imposed its own naming convention.
Geeqie is fair to do that job, just I'm looking to some feature to automate the process.

JackH79 · 2012-07-30 23:15:05

Use digikam.
Yes it's enormous, but it does most of what you want. It's working with a database, so multiple folders in various locations aren't a problem. And the compare algorithms are pretty good.

TheSaint · 2012-07-31 12:09:53

I forgot about digikam, I'll give it a try. Unfortunately it will ask a big chunk of KDE files, which I didn't install.
For Geeqie, above mentioned, I found that the collections is a plain text which contain the full path of each file inserted. I'll try to compose such list by the mean of python and see what will come out.

Rasi · 2012-07-31 20:05:34

I bloody hate digikams find duplicates function. You have to click every single file manually. no way to mass select the results, which is a major pain to do.

TheSaint · 2012-08-01 11:28:57

It took me to set up 153 Mb to get digikam running
But my research lead to findimagedupes. A tiny program to solve a part of my problems. I need some more studies to make successful results within small resources.
I'm running a netbook and I should have to find duplicates in a number of 6000 pictures.
I also seen some python scripts, but I think it'd need a small UI to see the results, for that I'm not so expert. Furthermore I don't need it more than a couple of time.

RE-EDIT

I used Geeqie and a small python script to compile a Geeqie collection.
The script takes 3 arguments:
SearchPic.py dir-to-scan "list, of, file, estensions" /file/to/save/the/results.gqv
Then from within the collection we can choose to find duplicates. A problem is about directories without permission, which makes scanning failure.

Last edited by TheSaint (2012-08-09 11:45:25)

Arch Linux

#1 2012-07-29 15:54:58

Duplicated image finder

#2 2012-07-29 16:25:51

Re: Duplicated image finder

#3 2012-07-29 16:35:21

Re: Duplicated image finder

#4 2012-07-29 16:38:40

Re: Duplicated image finder

#5 2012-07-30 10:12:45

Re: Duplicated image finder

#6 2012-07-30 23:15:05

Re: Duplicated image finder

#7 2012-07-31 12:09:53

Re: Duplicated image finder

#8 2012-07-31 20:05:34

Re: Duplicated image finder

#9 2012-08-01 11:28:57

Re: Duplicated image finder

Board footer