Digital Color Image Classification Based on Modified Local Binary Pattern Using Neural Network

Abstract

Classification of digital color images is one of the preeminent crucial processes in the image processing field since it increases the efficiency of the digital image recognition system. This paper presents a highly accurate and effective technique to extract color image features, which can be passed as an input data set of artificial neural network for classification purposes. The proposed method will be tested utilizing color images database. Then, the efficiency parameters will be determined in order to demonstrate the exactness and proficiency of the proposed technique. A modified local binary pattern method of extracting color image features, which can be used later as a signature to recognize color image is proposed and discussed. The proposed method is employed to create a feature array for each color image. The suggested classification technique using artificial neural network chops down the memory address space needed to store the necessary data required for the recognition system. Hence, the recognition time will be reduced.