On Fri, May 18, 2012 at 8:08 PM, David Warde-Farley <
[email protected]> wrote:
> On 2012-05-18, at 9:45 PM, [email protected] wrote:
>
> >
> > Here is sample problem:
> > I have cropped part from original image: http://i.imgur.com/nYYNf.pngthat
> > is part from original image:
> http://i.imgur.com/j2Lhb.png
> > Problem is, given the cropped part, find if it matches provided
> arbitrary image while resolution may not be same
>
> Look into SIFT features (patented, so be careful about commercial use) and
> other gradient histogram based keypoint descriptors. You don't actually
> need machine learning for this, a SIFT extractor and some sort of nearest
> neighbour lookup data structure will do. You might use PCA or something to
> reduce the dimensionality of the nearest neighbour problem, though.
check out opencv (http://code.opencv.org/projects/OpenCV/wiki/WikiStart).
it has full python bindings, and they may even have an example that does
exactly what you're asking about.
bryan
------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
Scikit-learn-general mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
