Proceedings of the IEEK Conference (대한전자공학회:학술대회논문집)

The Institute of Electronics and Information Engineers (대한전자공학회)
권호 표지

A new JPEG quantization table design for face recognition

얼굴인식을 위한 JPEG 양자화 테이블의 설계 방법

  • Ahn, Bong-Ju (Department of Electronic Engineering, Kookmin University) ;
  • Ka, Chung-Hee (Department of Electronic Engineering, Kookmin University) ;
  • Jeong, Gu-Min (Department of Electronic Engineering, Kookmin University) ;
  • Kim, Do-Hyun (Department of Electronic Engineering, Kookmin University)
  • Published : 2006.06.21
Download PDF
⟨ Previous Next ⟩

Abstract

In this paper, a codec design method is pro-posed for the face images based on JPEG and its application to face recognition is presented. Quantization table design is dis-cussed using R-D optimization for Yale face data. For the usage in the embedded systems, fast codec design is also considered. The proposed codec has better performance than JPEG codec for face images. Through the recognition experiment using PCA and LDA, it has been shown that the proposed codec has better performance than JPEG codec.

Keywords