I need matlab code for optimization techniques which is suited for texture image

I am working on texture image classification using BPNN classifier.I have extracted GLCM feature and I need to select the best features among 22 GLCM features.I need the best Optimization technique for texture images with code

[Walter Roberson]

What is being optimized?

[GNANAPRAKASH V]

Thanks for your comment sir. I need to optimize the GLCM parameter.

[Walter Roberson]

Which GLCM parameters? Number of intensity levels?

[GNANAPRAKASH V]

I have calculated 22 GLCM feature like contrast,correlation,entropy,homogeniety,mean,varaince,angular second moment and so on. Intensity level of an image is 0-255

[Image Analyst]

You're not answering clearly in a way that allows anyone to help you. OK, let's try it a different way. Let's say that you had two sets of GLCM parameters: one that was "optimized" and one that was NOT. How would you know the difference? Or, if I just gave you a random set of parameters, and you told me "No, that is not optimized", then how would you know to say that? What metric are you looking at to determine if the set of parameters is good or not?

[Walter Roberson]

Are you trying to do Feature Selection? https://www.mathworks.com/discovery/feature-selection.html

[GNANAPRAKASH V]

I have to select best features among 22 GLCM features and those selected features are given to the input of BPNN to find the performance metric like accuracy, sensitive and specificity

[Walter Roberson]

Then have a look at the above link, which outlines the available choices.

[GNANAPRAKASH V]

I tried with the above link.But I didn't get any source code here.May I get the source code for feature selection from redundant features.

[GNANAPRAKASH V]

I have calculated 22 GLCM features for 200 image data sets.I stored those features in EXCEL file.I have to import from EXCEL file to select the best features among redundant features.

[Walter Roberson]

That link refers to Regularization for removing redundant features. That was a kink to https://www.mathworks.com/discovery/regularization.html

That page in turn links to two blog posts and an example with complete code shown for them.

[GNANAPRAKASH V]

Can I import excel file data to that code....

[GNANAPRAKASH V]

Where I can do my novelty work in that code or feature selection algorithm

[Walter Roberson]

"Can I import excel file data to that code...."

No. But you can modify the code to accept data from a file.

"Where I can do my novelty work in that code or feature selection algorithm"

Anywhere you want.

I am getting the impression that you are looking for complete code that already does everything you need, and yet at the same time is very very obvious as to how you would modify it to do new things (and yet at the same time it has to already have those new things already in place.)

The material I linked to is blog posts and sample code and File Exchange Contributions and a Webinar, all demonstrating how to use the techniques. It is not intended as an exact solution for your needs. The exact code for what you need does not already exist, and you need to write it.

[GNANAPRAKASH V]

Thank you sir.