A Comparative Analysis of Feature Sets for Image Classification Using Back Propagation Neural Network

Abstract

The exponential growth in image data over the internet has resulted in a growing need for searching images according to our requirements. Content based image retrieval systems extract similar images from databases or the internet for facilitation of their users. A number of different feature sets and classifiers have been used by researchers for content based image retrieval. The goal of this research is to evaluate some common features sets used for classification of images and identify the best features depending upon the user requirement. Some commonly used features have been studies and a set of six feature sets have been selected for evaluation by the Back-Propagation Neural Network (BPNN). The results have been evaluated on the basis of precision and recall and it can be concluded that for natural images none of the feature sets perform well universally on all classes and the selection of optimal feature set depends on the type/class of images.

A New Approach of Content Based Image Retrieval Using Color and Texture Features

“…After that, in the year 2015, Husain and Akbar [16] proposed a method for retrieving images from the database using the classifier, Back Propagation Neural Network (BPNN) which consists of an input, a hidden and an output layer. In their method, they evaluated some common features sets to classify images, and identify the relevant features for the users by selecting 50 nodes for the hidden layer based on experiment.…”

An Efficient Classification Algorithms for Image Retrieval Based Color and Texture Features

“…Problems related to the image retrieval consist of two sections; (feature extraction and classification). Many different techniques and algorithms have been used by lot of researchers to extract image features and classification to improve recall and accuracy [3].…”