Hi everyone,
Some time ago I played around with computer vision to get images categorized on Commons. I documented this at https://commons.wikimedia.org/wiki/User:Multichill/Using_OpenCV_to_categoriz... . I don't think I'm going to spend time on it soon, but the results were quite promising, so maybe someone else feels like working on this? Would probably be a pretty nice student project or just fun to do.
Maarten
Hi Maarten;
I think that this is a perfect example of open question in wiki research. WikiPapers has a page for that stuff.[1] Can you add some bits there about this?
I dind't know about OpenCV, I will check it for sure, and I will try to something (I'm a bot developer).
Regards, emijrp
[1] http://wikipapers.referata.com/wiki/List_of_open_questions
2012/2/20 Maarten Dammers maarten@mdammers.nl
Hi everyone,
Some time ago I played around with computer vision to get images categorized on Commons. I documented this at https://commons.wikimedia.org/**wiki/User:Multichill/Using_** OpenCV_to_categorize_fileshttps://commons.wikimedia.org/wiki/User:Multichill/Using_OpenCV_to_categorize_files. I don't think I'm going to spend time on it soon, but the results were quite promising, so maybe someone else feels like working on this? Would probably be a pretty nice student project or just fun to do.
Maarten
______________________________**_________________ Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/**mailman/listinfo/wikitech-lhttps://lists.wikimedia.org/mailman/listinfo/wikitech-l
I have found a tutorial for Python coders http://creatingwithcode.com/howto/face-detection-in-static-images-with-pytho... some tests, it works fine (including René Descartes face : )).
This is going to be very helpful to improve Images for biographies accuracy http://toolserver.org/~emijrp/imagesforbio/
2012/2/20 emijrp emijrp@gmail.com
Hi Maarten;
I think that this is a perfect example of open question in wiki research. WikiPapers has a page for that stuff.[1] Can you add some bits there about this?
I dind't know about OpenCV, I will check it for sure, and I will try to something (I'm a bot developer).
Regards, emijrp
[1] http://wikipapers.referata.com/wiki/List_of_open_questions
2012/2/20 Maarten Dammers maarten@mdammers.nl
Hi everyone,
Some time ago I played around with computer vision to get images categorized on Commons. I documented this at https://commons.wikimedia.org/**wiki/User:Multichill/Using_** OpenCV_to_categorize_fileshttps://commons.wikimedia.org/wiki/User:Multichill/Using_OpenCV_to_categorize_files. I don't think I'm going to spend time on it soon, but the results were quite promising, so maybe someone else feels like working on this? Would probably be a pretty nice student project or just fun to do.
Maarten
______________________________**_________________ Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/**mailman/listinfo/wikitech-lhttps://lists.wikimedia.org/mailman/listinfo/wikitech-l
Maarten Dammers wrote:
Some time ago I played around with computer vision to get images categorized on Commons. I documented this at https://commons.wikimedia.org/wiki/User:Multichill/Using_OpenCV_to_categoriz... files . I don't think I'm going to spend time on it soon, but the results were quite promising, so maybe someone else feels like working on this? Would probably be a pretty nice student project or just fun to do.
Neat! :-)
I've had some thoughts about the uses of this. Particularly doing fuzzy image matching to spot duplicates, rather than relying on the primitive hash of the contents of the file. It might also be possible to do other fun searches, like looking for watermarks or poorly rotated photos (without the EXIF data, I guess) or making/breaking CAPTCHAs or...
I wonder if facial recognition will soon need to have some kind of opt-out.
MZMcBride
Hi MZMcBride,
Op 20-2-2012 23:24, MZMcBride schreef:
Maarten Dammers wrote:
Some time ago I played around with computer vision to get images categorized on Commons. I documented this at https://commons.wikimedia.org/wiki/User:Multichill/Using_OpenCV_to_categoriz... files . I don't think I'm going to spend time on it soon, but the results were quite promising, so maybe someone else feels like working on this? Would probably be a pretty nice student project or just fun to do.
Neat! :-)
I've had some thoughts about the uses of this. Particularly doing fuzzy image matching to spot duplicates, rather than relying on the primitive hash of the contents of the file.
If you want to find duplicate images you might want to play around with http://svn.wikimedia.org/viewvc/pywikipedia/trunk/pywikipedia/match_images.p...
Maarten
I looked at perceptual hashing some time ago. A problem was that it had to keep the uncompressed image in memory, and it was slow for being applied on commons scale.
Le 20/02/12 22:08, Maarten Dammers a écrit :
Some time ago I played around with computer vision to get images categorized on Commons. I documented this at https://commons.wikimedia.org/wiki/User:Multichill/Using_OpenCV_to_categoriz... . I don't think I'm going to spend time on it soon, but the results were quite promising, so maybe someone else feels like working on this? Would probably be a pretty nice student project or just fun to do.
Google used a game to add keywords to images. Two people (of of them could be a robot) were presented two pictures. The aim is to enter a keyword that best describe the picture and that the other people will choose too. If both people use the same keyword, you gain points :-)
Imagine, you are being shown a picture whose main subject is a monkey, you enter "monkey", the other one "monkey" too. You both earn points. The game host now know that the picture is of a monkey :-D
On Tue, Feb 21, 2012 at 8:09 AM, Antoine Musso hashar+wmf@free.fr wrote:
Le 20/02/12 22:08, Maarten Dammers a écrit :
Some time ago I played around with computer vision to get images categorized on Commons. I documented this at
https://commons.wikimedia.org/wiki/User:Multichill/Using_OpenCV_to_categoriz... . I don't think I'm going to spend time on it soon, but the results were quite promising, so maybe someone else feels like working on this? Would probably be a pretty nice student project or just fun to do.
Google used a game to add keywords to images. Two people (of of them could be a robot) were presented two pictures. The aim is to enter a keyword that best describe the picture and that the other people will choose too. If both people use the same keyword, you gain points :-)
Imagine, you are being shown a picture whose main subject is a monkey, you enter "monkey", the other one "monkey" too. You both earn points. The game host now know that the picture is of a monkey :-D
The "Google Image Labeler", there are (funny and offensive) videos of that online. It is a good idea. Mediawiki does not need a person to check it live, it has users that are reviewing the changes anyway.
It would be nice to have a *simple* way to put people down and have them label images, I could find people if it was really simple. Right now it is not really simple.
Most people just dont have the concentration to read the docs or learn how to use the wiki.
mike
wikitech-l@lists.wikimedia.org