Feature Extraction: Comparing Feature Extractors

Abstract

Machine learning is getting fuzz nowadays a lot. In order to train the computer to recognize the set of images is the key factor. For this neural networks and classifiers are used to do the identification and recognition. Now do the matching feature points or the interest points are required which makes it to our area of research i.e. feature extractors. In our paper we compared various feature extractors Fast, Eigen, Surf, Harris, Brisk and Mser. These were compared on the basis of number of feature points extracted, strongest feature points and the region in which these features are concentrated. For this procedure, MATLAB was used.