A wavelet based low complexity embedded block coding algorithm

Abstract

Summary form only given. Along with compression efficiency, other factors, like complexity, are significant issues for image coding. The measure of complexity varies from application to application. To overcome the problems of large database maintenance and the high computational burden of EZT and SPIHT, a new algorithm, WEBLOC (Wavelet-based Embedded BLOck Coding), is proposed for low complexity, near lossless compression. The most significant characteristics of this algorithm involve (a) sign-bit arrangement (b) subband intensity distribution statistics. Results were obtained with monochrome 8 bpp, 256/spl times/256 images. The salient features of the coding algorithm can be summarized as follows. Entropy coding is replaced by fixed RLC (runlength coding), considerably reducing the computational overhead and also the time overhead. For near lossless image compression, the reduction of complexity highly reduces the hardware circuitry. The running memory overhead for any list is reduced as compared to the EZT and SPIHT. However, the memory requirement for storing the wavelet coefficients is not reduced.