運用雙影像視覺與主要因素分析(PCA)於人臉辨識

Integrating Stereovision and Principle Component Analysis for Face Recognition

田方正(Fang-Chen Tien) ； 洪贊原(Zan-Yuan Hong) ； 蔡全益(Chuan-Yi Tsai) ； 田方治(Fang-Chih Tien)

《品質學報》 13卷1期 (2006/03) Pp. 11-19

摘要

本研究之目的在運用雙影像系統(Stereovision)建立一有效三維之臉部辨識系。首先，以雙影像系統取得之三維臉部資訊，再以主成份分析(PCA, principal component analysis)進行辨識的臉部辨識系統。雙影像視覺(Stereovision)技術是運用兩影像視差(disparity)特性藉以恢復物體之三維資訊，本研究之取40組臉部影像作實驗，每組影像取三種不同之表情。其中，以原始表情當訓練影像，其餘兩種表情為測試影像。結果顯示，以三維臉部資訊辨識率可達87.5%，故於小型之資料庫中可成功地辨識。

關鍵字

主要因素分析 ； 雙影像視覺 ； 視差 ； 臉部辨識

並列摘要

This paper presents a 3D face recognition technique developed based on the stereovision system. The proposed method first extracts the information of 3D face by means of a stereovision system that restores the 3D information of an object by deriving the disparity of two images, and then uses the principal component analysis algorithm to learn and recognize faces. This research digitizes 120 images, including three face images for each 40 persons, and uses one normal image for training and the other two images for testing. Experimental results show that 87.5% recognition rate is attained with the 3D face information.

並列關鍵字

Face Recognition ； Principle Component Analysis ； Stereovision

參考文獻

黃雅軒(2001)。淺談生物認證技術。電腦與通訊。96，65-73。

Achermann, X. Jiang,H. Bunke(1997).Face recognition using range images.International Conference on Virtual Systems and Multimedia.(International Conference on Virtual Systems and Multimedia).

Beumier,M. Acheroy(2000).Automatic face verification from 3D and grey level cues.Eleventh Portuguese Conference on Pattern Recognition.(Eleventh Portuguese Conference on Pattern Recognition).

B. Moreno,Angel S`anchez, J. F. V`elez,F. J.D`iaz(2003).Face recognition using 3D surface-extracted descriptors.Irish Machine Vision and Image Processing Conference (IMVIP 2003).(Irish Machine Vision and Image Processing Conference (IMVIP 2003)).

Chang, K. Bowyer,P. Flynn(2003).Face recognition using 2D and 3D facial data.(2003 Multimodal User Authentication Workshop).