Embedded to Lossless Image Coding (elic)

In this paper, a novel approach for Embedded to Lossless Image Coding (ELIC) is presented. With embedded to lossless coding, lossy encodings of the image at all lower bit rates are embedded in the lossless bit stream. This algorithm uses wavelet transforms that map integers to integers with embedded coding. The coding is based on bit significance criteria in which coefficients that become significant in each bit plane are identified using a significant switch mask. The significance switches are coded using adaptive quad tree splitting based on the prediction of a sub band from its previously coded ‘parent’ sub band. Context based adaptive arithmetic coding is used for both switching and data refinement. Usually prediction based lossless coding techniques outperform wavelet based lossless coding methods. The lossless performance of ELIC is comparable to both JPEG-LS, a non embedded algorithm, and SPIHT. For high bit rate coding at near lossless levels, ELIC shows a better performance than of SPIHT.