Title :
Large block RVQ with multipath searching
Author :
Kossentini, Faouzi ; Smith, Mark J T ; Barnes, C.F.
Author_Institution :
Georgia Tech Lorraine, Metz, France
Abstract :
An improved residual vector quantization (RVQ) method for image coding is presented. The RVQ structure has the desirable effect of reducing both computation and memory requirements. However, the performance of RVQ is not as good as that of exhaustive search VQ. Multipath searching is investigated with respect to the design algorithm and encoder implementation. Emphasis is given to the tradeoffs between complexity and performance of RVQ in the context of image coding with vectors of large dimension. Experimental results indicate that coded image quality can be improved by using this approach without significantly increasing complexity
Keywords :
image coding; vector quantisation; coded image quality; complexity; encoder implementation; image coding; large block quantisation; multipath searching; residual vector quantization; Algorithm design and analysis; Degradation; Design methodology; Image coding; Image quality; Image reconstruction; Rate-distortion; Speech coding; Tree data structures; Vector quantization;
Conference_Titel :
Circuits and Systems, 1992. ISCAS '92. Proceedings., 1992 IEEE International Symposium on
Conference_Location :
San Diego, CA
Print_ISBN :
0-7803-0593-0
DOI :
10.1109/ISCAS.1992.230505
