Face detection using discriminating feature analysis and support vector machine in video

Peichung Shih, Chengjun Liu

Research output: Chapter in Book/Report/Conference proceedingConference contribution

15 Scopus citations

Abstract

This paper presents a novel face detection method in video by using Discriminating Feature Analysis (DFA) and Support Vector Machine (SVM). Our method first incorporates temporal and skin color information to locate the field of interests. Then the face class is modelled using a small training set and the nonface class is defined by choosing nonface images that lie close to the face class. Finally, the SVM classifier together with Bayesian statistical analysis procedure applies the efficient features defined by DFA for face and nonface classification. Experiments using both still images and video streams show the feasibility of our new face detection method. In particular, when using 92 images (containing 282 faces) from the MIT-CMU test sets, our method achieves 98.2% correct face detection accuracy with 2 false detections. When using video streams, our method detects faces reliably with computational efficiency of more than 20 frames per second.

Original languageEnglish (US)
Title of host publicationProceedings of the 17th International Conference on Pattern Recognition, ICPR 2004
EditorsJ. Kittler, M. Petrou, M. Nixon
Pages407-410
Number of pages4
DOIs
StatePublished - Dec 17 2004
EventProceedings of the 17th International Conference on Pattern Recognition, ICPR 2004 - Cambridge, United Kingdom
Duration: Aug 23 2004Aug 26 2004

Publication series

NameProceedings - International Conference on Pattern Recognition
Volume2
ISSN (Print)1051-4651

Other

OtherProceedings of the 17th International Conference on Pattern Recognition, ICPR 2004
CountryUnited Kingdom
CityCambridge
Period8/23/048/26/04

All Science Journal Classification (ASJC) codes

  • Computer Vision and Pattern Recognition

Fingerprint Dive into the research topics of 'Face detection using discriminating feature analysis and support vector machine in video'. Together they form a unique fingerprint.

Cite this