I am implementing GLCM on MRI images. And also have got its features, & got the results too, but I don't understand the significance and utilization of all the features like entropy, contrast, etc. I know definition and equations, but want to know how to utilize all these features for classification and get the result.
each features like entropy, contrast, .... suitable for extraction of a certain land cover you should open each feature and from visual inspection you can know the most suitable feature for your target then you make layerstack of this feature with the panochromatic image or multispectral image then make the classification
the result will be better than classification of multispectral image only
it will improve the classification accuracy
In order to utilize the features you have obtained to classify. You need to identity the most important feature out of all the features you have calculated.
So basically what you did was feature extraction and now you have do feature selection. I applied the same concept on Sentinel-1A images. I calculated JM distance and Transform divergence score and based on the score, you can make various combinations of features you have obtained. This was you can see which feature is helpful in improving the accuracy.
This would be a great question to post in our new free medical imaging question and answer forum ( www.imagingQA.com ), there are a number of machined learning experts in the community. If useful, please feel free to open a new topic at the link below :
https://community.imagingqa.com/c/questions/5
I used the GLCM along with the LBP to classify vegetation density from RGB satellite images please reefer to this study: Article Detection of Vegetation Encroachment in Power Transmission L...