An Improved Neural Image Compression Approach with Cluster Based Preprocessing

The convergence time for training back propagation neural network for image compression is slow as compared to other traditional image compression techniques. This article proposes a pre-processing technique i.e. Pre-processed Back propagation neural image compression (PBN) with an enhancement in performance measures like better convergence time with respect to decoded picture quality and compression ratios as compared to simple back-propagation based image compression and other image coding techniques for color images.