人臉辨識在Android平台之實現

Face recognition realized in Android platform

廖志儒(Zhi-Ru Liao)

指導教授 ： 鍾鴻源

國立中央大學/資訊電機學院/電機工程學系/碩士(2014年)

摘要

本文目的在於開發應用程式在Android智慧型手機上應用，系統包含人臉偵測和人臉辨識兩部分。第一部分為人臉偵測，第二部分為人臉辨識，因為人臉偵測是人臉辨識的前置作業，其結果足以影響整個系統的效能，所以簡易人臉偵測和辨識，極為重要。由於本應用程式是即時系統，如何準確、快速地定位出人臉區域是開發辨識系統的主要目標。當影像輸入時，系統先利用顏色資訊從背景中分離出可能是人臉存在的膚色區塊，接著利用人臉區域中存在眼睛和嘴唇的區塊特徵，擷取出並改良之，然後再利用的人臉的幾何關係標定出正確的人臉位置，再以主成分分析的方法，簡化資料，再進行人臉辨識。

關鍵字

人臉偵測 ； 人臉辨識 ； Android

並列摘要

The purpose of this system is to develop an application on Android smartphone. The system consists of face detection and face recognition. The first part is to detect face. The second part is face recognition. Before face recognition, we start face detection. The efficient of face detection will affect the results of overall system performance. The system is a real-time system, the main factor of face detection is speed and accurate. First, the system separates the human face candidates from the background by color information when smartphone’s camera catch image. The system found the eyes and lip candidates from face candidates by feature methods. Next, the system locates the real face region using color information and the geometrical relation of eyes. Finally,the recognition process is verified by applying PCA algorithm.

並列關鍵字

face detection ； face recognition ； Android

參考文獻

[2]Sanjay Kr. Singh, D. S. Chauhan, Mayank Vatsa and Richa Singh,“A Robust Skin Color Based Face Detection Algorithm”, Tamkang Journal of Science and Engineering, pp. 227-234, 2003

[1]Sang-Hoon Kim and Hyoung-Gon Kim, “Face Detection using Multi-model Information”, IEEE Conference on Automatic Face and Gesture Recognition, pp. 14-19, 2000

[3]Kin Choong Yow and Roberto Cipolla, “Feature-based human face detection”, Image and Vision Computing, pp. 713-735, 1997

[4]C. Kotropoulos., “Rule-based face detection in frontal views”, IEEE Conference on Acoustics, Speech, and Signal Processing, pp. 2537- 2540, 1997

[5]C. Schneider, N. Esau, L. Kleinjohann and B. Kleinjohann, “Feature based Face Localization and Recognition on Mobile Devices”, IEEE Conference on Control, Automation, Robotics and Vision, pp.1-6, 2006