Title :
A new approximate Karhunen-Loeve transform for data compression
Author :
Pirooz, Ali D. ; Reed, Irving S.
Author_Institution :
Commun. Sci. Inst., Univ. of Southern California, Los Angeles, CA, USA
Abstract :
A new approximate Karhunen-Loeve transform (AKLT/sub 2/) is presented. The second-order perturbation of the eigenvectors of the first-order Markov sequence has been used in order to derive the AKLT/sub 2/. It is shown that the AKLT/sub 2/ is more efficient than the DCT and the AKLT in energy compaction and signal decorrelation.
Keywords :
Karhunen-Loeve transforms; Markov processes; approximation theory; covariance matrices; data compression; decorrelation; eigenvalues and eigenfunctions; inverse problems; DCT; approximate Karhunen-Loeve transform; auto-covariance matrix; data compression; eigenvalues; eigenvectors; energy compaction; first-order Markov sequence; inverse transform; second-order perturbation; signal decorrelation; Compaction; Covariance matrix; Data compression; Decorrelation; Discrete cosine transforms; Discrete transforms; Eigenvalues and eigenfunctions; Karhunen-Loeve transforms; Markov processes; Signal analysis;
Conference_Titel :
Signals, Systems & Computers, 1998. Conference Record of the Thirty-Second Asilomar Conference on
Conference_Location :
Pacific Grove, CA, USA
Print_ISBN :
0-7803-5148-7
DOI :
10.1109/ACSSC.1998.751571
