An efficient visual pattern block truncation coding

Abstract

This paper presents a new practical image coding system, visual pattern block truncation coding (VPBTC). The first idea of this new algorithm is the application of block truncation coding (BTC), which encodes an original image and obtains a bitmap. This bitmap is applied to computing the block gradient orientation and matching the block pattern. Another refinement is that the classification of blocks are based on properties of human visual perception. Blocks are classified into two types: uniform blocks and edge blocks. Not only can VPBTC remove the drawbacks of BTC (e.g. higher bit rate) but also can it repair the defects of VPIC (e.g. mismatching pattern). Four benchmarks prove that the MSE of this algorithm is less than that of VPIC by 37%. Its bitrate, furthermore, is about 60% of BTC's. The advantages of this coding method are fine edge preservation, low computational complexity, and easy implementation.