本研究的目的在於研究Android上人臉偵測方法，並且加以應用在實作上。首先，本研究應用Android偵測人臉的方法，是透過彈性束圖匹配法（Elastic Bunch Graph Matching簡稱EBGM）的技術可偵測出人臉之區塊。為使人臉區塊更加完整，本研究採用Android 4.0 SDK提供的API提供了一個彈性束圖匹配法，這兩個API分别是android.media.FaceDetector和android.media.FaceDetector.Face，應用於複雜背景、多重人臉之數位影像上，確實能快速且正確地偵測出多重人臉存在的區域，並能快速地標定出個別人臉之位置。最先使用人臉偵測的設備是數位相機，近年來智慧型手機越來越普及，尤其Android系統使用者的人數日趨增加，因此本文針對此系統的人臉偵測技術加以研究並且改善其功能，希望可以應用在點名系統上加以實作。
The purpose of this research is to look into the face detect method on Android system, and apply it practically in real world. First of all, through the research, we understand that EBGM (Elastic Bunch Graph Matching) is used on Android to detect the approximate features for human face. In order to make it more accurate and complete, two APIs of Android 4.0 SDK, android.media.FaceDetecter and android.media.FaceDetecter.Face, were used in this research to develop an application of EBGM. When applying to complex background and multiple faces in arbitrary image, the APIs can quickly and accurately detect the features for multiple faces as well as identify their exact location. The first device to use the face detection method is digital camera. Recently, smart phone has become very popular, especially the rapid increasing volume of Android phone users. Hence, this research is expected to improve the technique and functionality of face detection on the Android system, and maybe in the future, be able to apply to attendance system.
