Dear R.A.J.

Most of the researchers use magnitude feature from Gabor filters

Hi,

Gabor filter gives  set of strong response for locations of the target images that have structures in this given direction.

It act also as an edge detector.

http://www.cs.utah.edu/~arul/report/node13.html

 Gabor filter is à band pass . You can choice its parameters To extract many features. For example skin détection in face recognition process. Look To my papers you'l find more. 

One of the attempts to use Gabor filters is to extract a Gabor jet, which is a collection of responses from different Gabor filters at given location in the image. These Gabor jets can be seen as local texture descriptors, and they are used for various tasks, such as keypoint localization, object recognition, face recognition, ...

Traditionally, only the magnitudes of the Gabor wavelets are considered to be useful, as the phase varies rapidly with the offset location of the filter. However, when handled correctly, the phase can be very informative, as I have shown in:

http://publications.idiap.ch/index.php/publications/show/2419

Dear Rajeshwari Anand Janthakal,

Look the link. May be useful.

http://www3.ntu.edu.sg/home/adamskong/publication/PR_Gabor.pdf

http://www.dis.uniroma1.it/~visiope/Articoli/Texture/Gabor/2002ieee-tip.pdf

Regards, Shafagat

It is efficient method for extraction features but unfortunately most of extracted features are redundant features (duplicated features)

Regards

Ali 

The Gabor filter does not capture Mean, Standard Deviation, Root Mean Square, Mean Absolute, Skewness (if you want to capture these you could use a sliding window). The Gabor filter extract  magnitude at different orientation and scales. For example a magnitude at 0 degree orientation or at 90 degree orientation (or different scales).

https://stackoverflow.com/questions/20608458/gabor-feature-extraction

Boundary,Shape,texture

Pixel differences for detectin variations in different image scales