View previous topic :: View next topic |
Author |
Message |
avx Advocate
Joined: 21 Jun 2004 Posts: 2152
|
Posted: Sun Jan 27, 2013 3:54 am Post subject: identify similiar/same images by content? |
|
|
I've got a few hundred k of photos here, where some are pretty similar in what they show, maybe even identical. I need a way to identify them and I need a tool for that. Something like tineye does on the web, but for local usage.
Anyone can hint me to something? _________________ ++++++++++[>+++++++>++++++++++>+++>+<<<<-]>++.>+.+++++++..+++.>++.<<+++++++++++++++.>.+++.------.--------.>+.>. |
|
Back to top |
|
|
Knute Guru
Joined: 28 Jun 2004 Posts: 362 Location: Fargo, ND
|
Posted: Sun Jan 27, 2013 4:33 am Post subject: |
|
|
Check out app-misc/fdupes.
http://netdial.caribe.net/~adrian2/fdupes.html
HTH _________________ Knute
----------
The human race has one really effective weapon, and that is laughter.
-- Mark Twain
If you want proof of that, take a look at windows sometime. |
|
Back to top |
|
|
avx Advocate
Joined: 21 Jun 2004 Posts: 2152
|
Posted: Sun Jan 27, 2013 5:42 am Post subject: |
|
|
Thanks, but I already know that and it only finds exact (checksums) matches, not same contents in different files. F.e. I need to find something which can identify the same picture in different (f.e.) jpeg compressions. _________________ ++++++++++[>+++++++>++++++++++>+++>+<<<<-]>++.>+.+++++++..+++.>++.<<+++++++++++++++.>.+++.------.--------.>+.>. |
|
Back to top |
|
|
javeree Guru
Joined: 29 Jan 2006 Posts: 453
|
Posted: Mon Jan 28, 2013 2:24 pm Post subject: |
|
|
digikam has a 'find duplicates' tool that find the same or very similar pics.
(E.g. it would label two pictures of a person taken a second apart, as a duplicate, even though the head had moved a bit).
So you still need to inspect manually to see if it is really a duplicate or a very-near duplicate. However, it is still convenient to do. |
|
Back to top |
|
|
avx Advocate
Joined: 21 Jun 2004 Posts: 2152
|
Posted: Tue Jan 29, 2013 4:35 pm Post subject: |
|
|
javaree, thanks for your input. Seems to work quite well in identifying stuff, though it's pretty slow and brings a lot of KDE-crap.
A commandline solution would be nice, if someone has one. _________________ ++++++++++[>+++++++>++++++++++>+++>+<<<<-]>++.>+.+++++++..+++.>++.<<+++++++++++++++.>.+++.------.--------.>+.>. |
|
Back to top |
|
|
|