Extracting Facial Features with Partial Feature Templates

概要

論文の詳細を見る
Feature extraction is a frequent and fundamental problem in pattern recognition. Particularly, it is a key point in constructing a face recognition system or a facial expression analysis system. However, it is a difficult task because facial features vary with different illumination conditions, face styles and facial expressions.Some methods and algorithms have heen developed to extract facial features. Methods using image features are introduced in [1][2][3]. Since these features can not be used to express facial organ positions directly, further process with constraints of geometric model is necessary for facial organ extraction. The method of [4] can figure out the exact facial organ shapes, but the finalresults depend heavily upon the initial position guesses, which is difficult to use when the face position is unknown.
一般社団法人情報処理学会の論文
1993-09-27