Facial features extraction and tracking are crucial steps for many multimedia communication applications such as automated visual interpretation, human face recognition, and development of high quality model-based coding (e.g. MPEG-4) systems. Among different facial features, eyes and lip play an important role in either recognition process or teleconferencing applications. However, automatic human facial features detection is a difficult task due to different scale, rotation and translation of the features. This paper describes a computational approach for locating the eye position from a given eye window. The proposed algorithm extracts the eye in two stages. The iris first is extracted using a region-based energy minimization approach from the membership map generated by the spatial fuzzy clustering technique. In second two, control points are detected to locate the sclera and two parabolas are used to model the upper and lower eyelids. Satisfactory results have been achieved with the proposed method.
[1]
Alan L. Yuille,et al.
Feature extraction from faces using deformable templates
,
2004,
International Journal of Computer Vision.
[2]
Alan Wee-Chung Liew,et al.
Fuzzy image clustering incorporating spatial continuity
,
2000
.
[3]
Harry Wechsler,et al.
Detection of faces and facial landmarks using iconic filter banks
,
1997,
Pattern Recognit..
[4]
Tomaso A. Poggio,et al.
Example-Based Learning for View-Based Human Face Detection
,
1998,
IEEE Trans. Pattern Anal. Mach. Intell..
[5]
Hong Yan,et al.
Locating and extracting the eye in human face images
,
1996,
Pattern Recognit..