Face Recognition New Approach Based on Gradation Contour of Face Color

In this paper, we introduce a new technique to recognize face image based on the gradation contour of face color. The problems remain in face recognition issue is about lighting, expressions and poses. Feature based method is considered highly successful and quite economical, however this approach is very sensitive to the light, viewing angles and poses. To handle this problem, it is better to represent the face using the 3D models, although the cost is too expensive. Related to lighting, naturally people tend to recognize the face shape based on it. According to that, the authors tried a face recognition approach using contour gradation on the face color. Variables tested in this study was Threshold Contour error, X error and Size error, then those variables are tested against the image with light illumination, yaw face and variation amount of contour lines. After testing, the best recognition results generated by 85.458% for Aberdeen face database with 150 contour lines and 90% for Yale face database. The other result showed that image with different expression was still recognizable however it cannot be recognized more than 20 degrees yaw faces. It indicated that the face contour as a feature was fairly representative to be used in face recognition.