An Approach to Color Image Coding Based on Adaptive Multilevel Block Truncation Coding

Abstract

Block Truncation Coding (BTC) has been considered as a prime choice in many instances because it is the simplest to implement and is quite satisfactory from the viewpoint of the resulting fidelity and compression ratio. To effectively compute multidimensional color data, an efficient expression of color data is necessary, therefore this paper presents a novel BTC algorithm called Adaptive Multilevel Block Truncation Coding (AMBTC) to achieve the optimal coding performance by adaptively choosing two-level or four-level BTC according to the edge property of the coding block. The AMBTC uses an adaptive selector level based on the sum of absolute difference (SAD) to get optimal coding performance. In order to reduce the bit rate, we utilize luminance bitmap that represents the three color bitmaps. This method improves the two-level and four-level BTC used in AMBTC. The simulation result is compared with existing conventional BTC and it is found that the proposed BTC outperforms in PSNR and compression ratio and achieves better image quality of reconstructed images.