Title :
Interpolation approaches to vector quantization for image compression
Author :
Carrato, Sergio ; Passaggio, Filippo ; Rovetta, Stefano ; Zunino, Rodolfo
Author_Institution :
Dipartimento di Elettrotecnica, Elettronica ed Inf., Trieste Univ., Italy
Abstract :
Vector quantization (VQ) applies effectively to very low bit-rate image transmission. The high compression ratios are often balanced by limited reconstruction quality due to blockiness effects due to coarse quantization. Interpolating different codevectors can overcome such drawback by enhancing the generalization ability and the adaptiveness of the coding system at run time. Experimental results on real testbeds show the method´s advantages as compared with related techniques
Keywords :
image coding; interpolation; vector quantisation; adaptiveness; blockiness effects; coarse quantization; codevectors; generalization ability; image compression; limited reconstruction quality; vector quantization; very low bit-rate image transmission; Image coding; Image communication; Image reconstruction; Interpolation; Iterative algorithms; Partitioning algorithms; Prototypes; Testing; Topology; Vector quantization;
Conference_Titel :
Neural Networks for Signal Processing [1996] VI. Proceedings of the 1996 IEEE Signal Processing Society Workshop
Conference_Location :
Kyoto
Print_ISBN :
0-7803-3550-3
DOI :
10.1109/NNSP.1996.548369
