Translator Disclaimer
3 March 1995 Experiments in lossless and virtually lossless image-compression algorithms
Author Affiliations +
Proceedings Volume 2418, Still-Image Compression; (1995)
Event: IS&T/SPIE's Symposium on Electronic Imaging: Science and Technology, 1995, San Jose, CA, United States
The CB9 lossless image compression algorithm is context-based, and codes prediction errors with an adaptive arithmetic code. It has been developed within an algorithm class that includes (in the order of their development) Sunset, JPEG lossless, sub8xb, and now CaTH (Centering and Tail Handling). Lossless compression algorithms using prediction errors are easily modified to introduce a small loss through quantization so that the absolute error for any pixel location does not exceed prescribed value N. In this case, N varies from 1 to 7; the values for which the JPEG group issued a call for contributions. This work describes CB9 and the experiments with near-lossless compression using the JPEG test images. Included are experiments with some image processing operations such as edge-enhancement with the purpose of studying the loss in fidelity from successively performing decompression, followed by an image processing operation, followed by recompression of the new result.
© (1995) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Glen G. Langdon Jr. and Chris A. Haidinyak "Experiments in lossless and virtually lossless image-compression algorithms", Proc. SPIE 2418, Still-Image Compression, (3 March 1995);


Recursive scaled DCT
Proceedings of SPIE (December 01 1991)
Enhancement of transform coding by nonlinear interpolation
Proceedings of SPIE (November 01 1991)

Back to Top