Title :
Separable FIR principal component filter banks
Author :
Xuan, Bo ; Bamberger, Roberto H.
Author_Institution :
Dept. of Electr. Eng. & Comput. Sci., Washington State Univ., Pullman, WA, USA
Abstract :
Principal component filter banks (PCFBs) minimise the mean squared error between the input and output signals when only a limited number of subbands are retained for reconstruction, hence attain the most energy compaction. It is proven that a PCFB maximizes the theoretical-coding gain of the uniform multi-rate analysis/synthesis paraunitary filter bank system. However, the PCFBs are typically composed of infinite impulse response with no simple closed forms. It is not practical to implement them. This paper discusses one particular type of 2D FIR PCFBs, the separable FIR PCFB, which has low design and implementation costs
Keywords :
FIR filters; adaptive filters; adaptive signal processing; band-pass filters; filtering theory; image reconstruction; two-dimensional digital filters; 2D FIR PCFB; coding gain; energy compaction; image processing; infinite impulse response; input signals; low design costs; low implementation costs; mean squared error minimisation; output signals; paraunitary filter bank; separable FIR principal component filter banks; signal reconstruction; uniform multirate analysis/synthesis filter bank; Autocorrelation; Channel bank filters; Compaction; Computer errors; Computer science; Filter bank; Finite impulse response filter; IIR filters; Signal analysis; Signal synthesis;
Conference_Titel :
Signal Processing, 1996., 3rd International Conference on
Conference_Location :
Beijing
Print_ISBN :
0-7803-2912-0
DOI :
10.1109/ICSIGP.1996.566969
