How to find features of an image?

How can i find features (Energy, entropy, contrast. correlation homogeneity etc)of an image. I searched in MATLAB Help.

But in that, only equations to calculate these features are given, but no direct commands.

My project is to find level of Diabetic Retinopathy disease using fundus (eye)images. I am done with two steps of my work: Pre-processing of images. Segmentation of images

Now the third step is to find features followed by Classification.

Can anyone help me in this.....

Answer 1

[Image Analyst]

Your first step is to look at section 20.5 here to see how others have done it: http://www.visionbib.com/bibliography/contentsmedical.html#Medical%20Applications,%20CAT,%20MRI,%20Ultrasound,%20Heart%20Models,%20Brain%20Models Next you have to decide if your job is to invent a new method or apply one of the exiting methods . If you need to invent a new or improved method (like you're working on your Masters or Ph.D. degree), after knowing all the other methods, you have to decide on your approach that would be new and novel, or at the very least, an improvement over existing methods. If you need to invent or improve a method, then you can't really ask us to develop your feature finding algorithm for you since it has to be yours, not ours. I can tell you that there are functions for the operations you listed (entropy, correlation, etc.) in toolboxes such as the Image Processing Toolbox. And I can also refer you to my File Exchange http://www.mathworks.com/matlabcentral/fileexchange/?term=authorid%3A31862 which has demos on grayscale and color segmentation methods. They will be a great start for you to learn the basics of how to do segmentation in MATLAB.

[Preeti Mistry]

Thank you.... The links that you send were really useful...

Thanks

Answer 2

[VALARMATHY K]

you can use this for your calculation

GLCM = [0 1 2 3;1 1 2 3;1 0 2 0;0 0 0 3];

stats = graycoprops(GLCM)

I = imread('cameraman.tif');

GLCM2 = graycomatrix(I,'Offset',[2 0;0 2]);

stats = graycoprops(GLCM2,{'contrast','homogeneity'})