顔らしさの信頼性に基づく自然背景からの複数顔領域の抽出

概要

論文の詳細を見る
In general, it is very difficult to extract human faces from a natural background. In conventional methods, several conditions, such as the number of people included in a picture, the size of a face or the texture of the background are assumed. In this paper, the extracting method of faces from more than one person’s facial area with a natural background is presented, however, in the proposed method, such conditions described above are not supposed. In this system, a facial area is extracted using only the information of skin color at the first frame. After the second frame, the facial area is extracted based on the moving information of a face in addition to the color information.Computer simulation using 605 frames shows more than 85% extracting accuracy. From the experimental results, the prospects of using this method for the human face identification, image coding or the recognition of facial expressions are very encouraging.
社団法人電気学会の論文
2003-03-01