Optimal design of transform coders and quantizers for image classification

S. Jana, P. Moulin

Research output: Contribution to conferencePaper

Abstract

In a variety of applications (including automatic target recognition) image classification algorithms operate on compressed image data. This paper explores the design of optimal transform coders and scalar quantizers using Chernoff bounds on probability of misclassification as the measure of classification accuracy. This design improves classification performance but the mean square error (as well as the visual quality) of the coded image degrades. However, by appropriately combining classification accuracy and mean square error in the cost function, one can achieve good classification with low (visual) distortion, which is desirable in classification systems requiring visual authentication.

Original languageEnglish (US)
Pages[d]841-844
StatePublished - Dec 1 2000
EventInternational Conference on Image Processing (ICIP 2000) - Vancouver, BC, Canada
Duration: Sep 10 2000Sep 13 2000

Other

OtherInternational Conference on Image Processing (ICIP 2000)
CountryCanada
CityVancouver, BC
Period9/10/009/13/00

ASJC Scopus subject areas

  • Hardware and Architecture
  • Computer Vision and Pattern Recognition
  • Electrical and Electronic Engineering

Fingerprint Dive into the research topics of 'Optimal design of transform coders and quantizers for image classification'. Together they form a unique fingerprint.

  • Cite this

    Jana, S., & Moulin, P. (2000). Optimal design of transform coders and quantizers for image classification. [d]841-844. Paper presented at International Conference on Image Processing (ICIP 2000), Vancouver, BC, Canada.