Compression of old marathi manuscript images using context-based, adaptive, lossless image coding

Abstract

Lossless compression ensures high computational and coding efficiency with lower model cost. The prediction and residual approach is commonly used to achieve this goal. Context-based Adaptive Lossless Image Coding is abbreviated as CALIC. This proves an efficient scheme of compression for continuous-tone images. The high coding efficiency is achieved in this scheme with relatively low space and time complexity. It uses simple and non linear gradient based prediction scheme GAP. Large numbers of modeling context are used to shape non linear predictor which makes it adaptive through error feedback system. CALIC scheme is used to estimate the expectation of prediction errors which is conditioned on large number of model context. It does not suffer from ‘context dilution’ problem. The core theme of CALIC is discussed here. Compression results of old Marathi manuscript test images prove superior performance of the CALIC compared with predictive Huffman and Arithmetic techniques implemented during experimentation.