Facial feature extraction and image-based face drawing

Abstract

This paper describes algorithms to extract features of facial organs such as eyes, eyebrows, mouth, nose and borders of other parts. Based on these extracted organs, automatic face drawing is generated, which embodies the individual features that the face looks more likely to be so. At first face detection is implemented by using horizontal textures and eigenfaces. After that, the pair of irises are located using the Hough-transform and intensity valley correlation. Then other organs are extracted based on the skeleton characteristic of the face. For face drawing generating, contours are rendered with multi-values, facial skin color is smoothed, and features are strengthened in order to get impressive effects.