Title :
Design of non-linear multi-resolution decompositions with applications to image compression
Author :
Cardoso, Eduarclo S., Jr. ; da Silva, E.A.B.
Author_Institution :
PEE/COPPE/DEL/EE, Univ. Fed. do Rio de Janeiro, Brazil
Abstract :
This paper suggests two novel heuristics to design maximally-decimated non-linear decompositions based on the lifting scheme. The first one dictates the desirable characteristics a nonlinear operator should have in order to be able to generate good decompositions. Based on this heuristic, we introduce two nonlinear operators using mathematical morphology. The second heuristic is related to the design of good nonlinear decompositions using the introduced operators in a lifting scheme. Simulation results show that nonlinear decompositions thus generated perform well in image coding schemes
Keywords :
data compression; image coding; image resolution; mathematical morphology; image coding schemes; image compression; lifting scheme; mathematical morphology; nonlinear multi-resolution decompositions; Character generation; DC generators; Image coding; Image reconstruction; Morphology; Signal generators; Smoothing methods; Tail;
Conference_Titel :
Circuits and Systems, 2001. ISCAS 2001. The 2001 IEEE International Symposium on
Conference_Location :
Sydney, NSW
Print_ISBN :
0-7803-6685-9
DOI :
10.1109/ISCAS.2001.921115
