Rate-distortion bound for joint compression and classification with application to multiaspect scattering

Yanting Dong, Shaorong Chang, Lawrence Carin

Research output: Contribution to journalArticlepeer-review

5 Scopus citations

Abstract

Rate-distortion analysis is applied to the problem of joint compression and classification. A Lagrangian distortion measure is used to consider both the Euclidean error in reconstructing the original data as well as the classification performance. The bound is calculated based on an alternating-minimization procedure, representing an extension of the Blahut-Arimoto algorithm. This rate-distortion framework is then applied to a joint compression and target-orientation estimation problem, based on a sequence of scattered waveforms measured at multiple target-sensor orientations. A hidden Markov model-Markov model (HMM-MM) is used as the statistical description for the source, here representative of multiaspect scattering data. Target-orientation estimation reduces to assessing the underlying HMM states from a sequence of observations. After deriving the rate-distortion function, we demonstrate that discrete HMM performance based on Lloyd encoding is far from this bound. Performance is improved via block coding, based on Bayes vector quantization. Results are presented for multiaspect acoustic scattering from an underwater elastic target, using measured and synthesized data. © 2005 IEEE.
Original languageEnglish (US)
Pages (from-to)481-492
Number of pages12
JournalIEEE Sensors Journal
Volume5
Issue number3
DOIs
StatePublished - Jun 1 2005
Externally publishedYes

Fingerprint Dive into the research topics of 'Rate-distortion bound for joint compression and classification with application to multiaspect scattering'. Together they form a unique fingerprint.

Cite this