How to do feature extraction using pixel value of an image?

Hello friends, In my project I want to extract features from an image. Input to the program is normal .jpg image while output is expected to be tabular form of number of pixels present in the image Example- If I is 8*8 image then output should show Number of Pixels- Red- xx Green- xx Blue- xx . . . and so on for large range of colors.

Answer 1

[Image Analyst]

See my gray scale image segmentation tutorial and color segmentation demos on my File Exchange: http://www.mathworks.com/matlabcentral/fileexchange/?term=authorid%3A31862, http://blogs.mathworks.com/pick/2009/11/06/segmenting-coinsa-tutorial-on-blob-analysis/

You only have 64 pixels in an 8 by 8 image, so what is "a large range of colors"? Each pixel will have a red, a green, and a blue value so I'm not sure what your output should show. Do you want to specify a certain number of colors, like 4 colors like you'd do with rgb2ind() and then give area fractions for each of the N color classes? Or do you want to do color classification like this: http://www.mathworks.com/products/demos/image/color_seg_k/ipexhistology.html

For further help: http://www.mathworks.com/matlabcentral/answers/7924-where-can-i-upload-images-and-files-for-use-on-matlab-answers and this: http://www.mathworks.com/matlabcentral/answers/6200-tutorial-how-to-ask-a-question-on-answers-and-get-a-fast-answer