how to interpret the output of gabor filter for texture segmentation?

Hi,

I need to segement texture but i dont know how to interpret the output image of the gabor filter.

this is the input image:

https://www.dropbox.com/s/2kctcbb9i4huqs0/input%20image.tif

this is the output image after I applied a gabor filter:

https://www.dropbox.com/s/gmn0pjbd45uwfxt/output%20image.jpg

if someone can explain me how to interpret it.

Answer 1

[Image Analyst]

It's like an edge detection filter, as explained here in Wikipedia: http://en.wikipedia.org/wiki/Gabor_filter. Basically if there is a high variation from one side of the sliding window to the other (like happens with an edge) then there is a high output signal. That's why the edges of the banana are bright in the output and the background is dark.

[Pablo]

Thanks for the answer it is very helpful because I need to detect the edges too. However I have researched it is used to segment texture too, but I don't know how they process the output image. I'm new in texture segmentation.

[Image Analyst]

I don't know what you want to do, but if you want advice, upload an image somewhere, such as http://snag.gy

[Pablo]

ok, thanks for the advice.

Answer 2

[Dakshayani Virodhi]

How to justify that the output of Gabor filter applied to the input image is correct or not?

[Image Analyst]

If you want you could pick a few pixels and calculate the result by hand and see if it agrees with the MATLAB function results. But I'm very sure that the Mathworks validated the math before they released the function so you can trust that it is correct.