自然屋内環境での表情認識を目的とした高精細な顔領域抽出

概要

論文の詳細を見る
An extraction method for facial areas taking into consideration pictures taken from any angles for a the backgroundenvironmental for facial expression recognition is presented in this paper. In a previous method, we used moving informationand color information. However, there still remain problems with several thresholds used for extracting facialareas. The threshold value was set as the fixed value. In the proposed system, in order to extract the moving area, theinter frame subtraction method with variable frame interval was adopted and an adequate frame interval was set. Moreover,by removing the neck region that is not required for facial expression recognition, the extracting accuracy of facescan be improved. From a computer simulation using 34278 images, 33366 images (97.34%) were correctly extracted.
2012-04-30