Research Article
Error correcting arithmetic coding for JPEG 2000: memory and performance analysis
@INPROCEEDINGS{10.1145/1374296.1374302, author={Simone Zezza and Marco Grangetto and Maurizio Martina and Fabrizio Vacca and Guido Masera }, title={Error correcting arithmetic coding for JPEG 2000: memory and performance analysis}, proceedings={2nd International ICST Conference on Mobile Multimedia Communications}, publisher={ACM}, proceedings_a={MOBIMEDIA}, year={2006}, month={9}, keywords={}, doi={10.1145/1374296.1374302} }- Simone Zezza
Marco Grangetto
Maurizio Martina
Fabrizio Vacca
Guido Masera
Year: 2006
Error correcting arithmetic coding for JPEG 2000: memory and performance analysis
MOBIMEDIA
ACM
DOI: 10.1145/1374296.1374302
Abstract
Entropy coding, and specifically arithmetic codes are particularly sensitive to bit errors. Indeed, due to the memory inherent to the arithmetic coding a single flipped bit may cause desynchronization of the decoder, hence all the remaining symbols can be erroneous. In this paper a detailed study of complexity, for a known in literature correction algorithm, is presented. In particular the amount of memory required to implement the algorithm has been evaluated, proposing two possible approaches: previous state and repeat current pass. Experimental results show that the second method requires a lower amount of memory than the first. Furthermore the performance achieved by the error correction algorithm, in terms of decoded frames per second, have been analyzed. To this purpose the MQ decoder rate has been evaluated.