A Novel Image-Compression Technique Using the Modular Algorithm

Jidong Shen

In predictive image coding, information loss presents a problem in accurately reconstructing the input image, because the probability distribution at different levels of the image-quantizer output is highly nonuniform and thus a huge codebook is required. This article presents a new technique for DPCM image compression, based on the modular algorithm. By using this technique, the length of codewords required by the entropy coding is necessarily limited to a certain positive integer, and with a smaller codebook we can distortionlessly reconstruct the input image from predictive decoding.

Print ISSN
Published
1993-05
Content type
Original Research
DOI
10.5594/J15914