Machine Learning Based Object Recognition

Abstract

In this paper Object Recognition technique is proposed, that produce the best idea to recognize the object from the given input image and it is implemented in MATLAB tool. The past few years, SVM has been applied and estimated only as pixel-based image classifiers. Recently pixel based process moving towards object recognition technique. In an analysis, the SVMs performances are compared with some other classifiers such that BPN classifier and KNN classifier. This technique is obtained by extracting the energies from wave atom transform. The extracted features are given to the SVM classifier as an input and recognize the corresponding image in an object. Finally the experimental results are shown for COIL-100 database. The result of our proposed method is evaluated to increasing the rate of recognition accuracy and correct recognition rate.