基於浮動參數與生物特徵的人臉特徵偵測

Facial Features Detection based on Floating Parameter and Biometrics

楊凱麟(Kai-Lin Yang)

指導教授 ： 陳榮昌

朝陽科技大學/資訊學院/資訊管理系/碩士(2013年)

摘要

人臉偵測是藉由人臉視覺特徵進行偵測的一種電腦技術，它被廣泛的應用在開車專心偵測、門禁系統及表情偵測等，經過人臉偵測之後就可以進行人臉辨識。人臉辨識是生物特徵辨識的一種技術，生物特徵包括臉、指紋、虹膜、及聲音等。由於現今資訊科技快速進步，個人隱私以及資料的保護也漸漸受到重視，傳統的身分辨識如密碼、識別卡、或鑰匙等，都容易偽造、遺失或是遭到破解。因為人臉偵測有獨一無二的特性，所以使用生物特徵辨識是目前較為方便和安全的辨識技術。但在人臉偵測與辨識有許多的困難會使得準確率偏低，如何提高準確率為本篇論文所探討的主題。
在本研究使用了基於特徵的人臉偵測方法結合浮動參數(Floating Parameter)設定，改善基於特徵的人臉偵測所會產生之缺點，並會檢查YCbCr的範圍來決定該區域是否為膚色。由於膚色範圍的設定如果能因環境的顏色而改變，將更能彈性的適應各種影像。另外，因為人臉特徵的大小會隨著輸入的圖片大小而不同，所以如果使用固定的特徵參數，將較難有效的辨識。因此所以本研究透過浮動參數運用於人臉偵測的實驗，以提升在不同的條件之下的人臉偵測準確率。

關鍵字

生物辨識 ； YCbCr ； 人臉偵測 ； 人臉辨識 ； 浮動參數

並列摘要

Face detection is a technology for detecting features of human face. It has a wide area of applications, such as driver gaze tracking accessing control system and facial expression recognition, etc. Face detection is an important stage before face recognition. Face detection is a technology in Biometrics (including face, finger point, iris, voice, etc.). Due to rapid progress in information technology, personal privacy and data protection are concerned. Traditional identity recognition is such as password, ID card, key, etc. They are easy to fake, lost or jailbreak. During the uniqueness property, Biometrics is the most convenient and safe technologies. But the methods of facial detection and recognition have many problems and have low accuracy, so how to improve accuracy are the main objectives of this study.
In this thesis, the face detection based on feature method combine floating parameter is used to improve the defect with face detection based feature method. The values of Cb and Cr are checked to find the skin color by skin color segmentation of YCbCr, the accuracy will be influenced by the brightness of the image. If the accepted ranges of Cb and Cr could be floating values which are dependent with environmental brightness, it will be more flexible than fixed ranges to adapt the varieties of images. So we research face detection with float parameter in different conditions, and to improve the accuracy.

並列關鍵字

YCbCr. ； Biometrics ； Floating Parameter ； Face Recognition ； Face Detection

參考文獻

[29] 薛傑仁，「生物辨識之人臉辨識的方法」，碩士論文，生物與醫學資訊系，亞洲大學，2010。

[24] 徐民儕，高文忠，「應用區域對比增強於不均勻光源下之人臉辨識」，應用電子科技研究所，國立臺灣師範大學，2008。

[5] S. Dutta and B. B. Chaudhuri, “A Color Edge Detection Algorithm in RGB Color Space,” International Conference on Advances in Recent Technologies in Communication and Computing, pp. 337-340, 2009.

[1] M.A. Berbar, H.M. Kelash, and A.A. Kandeel, “Faces and Facial Features Detection in Color Images,” Geometric Modeling and Imaging, pp. 209-214, 2006.

[2] A.P. Chen, L. Pan, Y.B. Tong, and N. Ning, “Face Detection Technology Based on Skin Color Segmentation and Template Matching,” Second Interactional Workshop on Education Technology and Computer Science, pp. 708-711, 2010.