Image Compression with Back Propagation Combined with Histogram Equalization

Abstract

This paper focuses on key research going in Neural Networks arena now days. It brings out its utility in area of Image compression .Image compression is not a novice concept. It has been in research for many years but continuously invites the improved techniques, for carrying out the compression effectively and efficiently. Role of Artificial Intelligence is being probe in area of compression and various neural network architectures are explored for Image compression. Basic Back propagation neural network is one among them. It employs simple three layer concept of: Input, Hidden and Output layer to perform compression but the loophole appears in it, in the form of Convergence time. Back propagation took longer time to compress because of its long process of weight adjustment. To circumvent this problem, Image preprocessing technique: Histogram Equalization is employed, which improves the No. of Epochs and thus the time required for compression.

Keywords : Histogram Equalization, Image Compression, Back Propagation, Image Processing.

View PDF