how to find mean absolute difference between two blocks of different images???

sir, first i want to find out the displacement between two similar blocks of different image frames. for that i need to compare the blocks of two images , to find the best matching block usine mean absolute difference(MAD)... what i have to do???

Answer 1

[Image Analyst]

There is also a Median Absolute Deviation which is more robust to outliers than the Mean absolute deviation. I have a demo for that it you want. Otherwise

madBlock = mean2(abs(double(block1) - double(block2)));

For color images, you'd have to do it for each color channel because mean2() works on 2D images, not 3D color images.

madBlockR = mean2(abs(double(block1R) - double(block2R)));
madBlockG = mean2(abs(double(block1G) - double(block2G)));
madBlockB = mean2(abs(double(block1B) - double(block2B)));

And there is the FAQ for dividing an image into blocks (two different ways): http://matlab.wikia.com/wiki/FAQ#How_do_I_split_an_image_into_non-overlapping_blocks.3F

Answer 2

[Nitin]

You can find a nice implementation of the above here