I solved this problem. I am not expecting an answer to this question. If any of you face a similar situation for them I am explaining here. Let you have texture features of two classes. Classify those features based on two/multiple different classifiers, for example, SVM, KNN, RF, Adaboost, etc. Compute probability scores using each classifier. Now, concatenate the probability scores of all classifiers and finally again classify the using any classifiers. In this step, you may use previously used classifiers.
CLICK HERE to find out more related problems solutions.