An efficient BTC image compression technique

This paper presents a moment preserving and visual information dominance technique to achieve the low-bit rate block truncation coding (BTC). Compared with other existing strategies as transform coding and vector quantization, conventional BTC compression has the advantage of simple and fast computation. Nevertheless the compression ratio is limited by its low efficiency. Our proposed technique accomplishes the goal of simple computation with variable bit rate selection by the moment preservation and information extraction algorithm. The proposed technique has the advantage of simple operations and it does not require complicated mathematical computations. Thus, the overall computation does not increase the burden compared with ordinary BTC. The simulations are carried with natural images to evaluate the performance. The generated decoded images have moderate quality with a bit rate of 0.5-1.0 bit/pixel.