Title :
Predictive coding using noncausal models
Author :
Balram, Nikhil ; Moura, José M F
Author_Institution :
IBM Corp., Boca Raton, FL, USA
Abstract :
A novel image compression algorithm is presented. The algorithm combines the recursive processing characteristic of predictive coding with the use of a noncausal image model. Experimental results demonstrate the high quality of the reconstructed image at a low bit rate (0.375 bit/pixel). This contrasts with the significant loss of detail and blocking artifacts introduced by a JPEG (Joint Photographic Experts Group) type DCT (discrete cosine transform) method at the same bit rate.<>
Keywords :
code standards; data compression; image coding; recursive functions; JPEG; blocking artifacts; discrete cosine transform; image compression algorithm; loss of detail; noncausal image model; predictive coding; recursive processing;
Conference_Titel :
Acoustics, Speech, and Signal Processing, 1993. ICASSP-93., 1993 IEEE International Conference on
Conference_Location :
Minneapolis, MN, USA
Print_ISBN :
0-7803-7402-9
DOI :
10.1109/ICASSP.1993.319888
