本文提出一套在複雜背景中自動將臉孔正確定位的方法, 由於複雜的背景嚴重的影響到人臉辨識的成果, 因此我們先對影像做邊緣偵測, 以加強臉部輪廓與背景之間的差異性。接著我們根據人臉輪廓與五官的對稱特性, 找出臉部在影像中的正確位置, 並針對原來影像的這個部分進行臉孔之定位工作。實驗結果顯示此技術在複雜背景之下都有良好的辨識效果。 In this paper, an automatic face location algorithm will be proposed. Detecting a face in a complex background is a difficult problem. We want to increase the difference between the face edge and the background by edge detection. The algorithm is based on the symmetry of a face and the relative position between hair and eye. Therefore the proposed approach can find the true position of human face in the image and the face region can be located and clipped out of the image. The experimental results show that the proposed algorithm for detecting face in a complex background is accurate and efficient.
一九九八年全國電信研討會論文集=Proceedings of 1998 National Symposium on Telecommunications，頁175-179