what algorithm good for image matching?

what algorithm 开发者_如何学编程good for image matching?to match different between two image

Here is a good Matlab reference for some image Computer Vision algorithms in Matlab that I've found useful in the past:

http://www.csse.uwa.edu.au/~pk/Research/MatlabFns/

The question is a little vague, but I would say the "sum of absolute differences" (SAD) might be what you're looking for.

http://en.wikipedia.org/wiki/Sum_of_absolute_differences

It's reasonably easy to understand/implement.

There are different good methods you can search and use:

1- HOG (Histogram of oriented gradients) 2- Hamming 3- SIFT

also check this library: VLFeat

There are a number of functions in FileExchange you can download and use freely.

One of the best computer vision techniques is to extract keypoints using SURF in the two images and then match the retrieved keypoints.

Regards,

OpenSurf is good for image matching as it can match image irrespective of skew or rotation. It simply calculates points in an image. To check if two images are same you just have to match the points and set a suitable ratio of match criteria.

https://code.google.com/p/opensurf1/downloads/detail?name=OpenSURFcs.zip&can=2&q=