Pose-Invariant Face Detection Using Edge-Like Blob Map and Fuzzy Logic

We present an effective method of face and facial feature detection under pose variation in cluttered background. Our approach is flexible to both color and gray facial images and is also feasible for detecting facial features in quasi real-time. Based on the characteristics of neighborhood area of facial features, a new directional template for the facial feature is defined. By applying this template to the input facial image, novel edge-like blob map (EBM) with multiple strength intensity is constructed. And we propose an effective pose estimator using fuzzy logic and a simple PCA method. Combining these methods, robust face localization is achieved for face recognition in mobile robots. Experimental results using various color and gray images prove accuracy and usefulness of the proposed algorithm. This research was supported by Korea Ministry of Science and Technology under the National Research Laboratory project, by Korea Ministry of Education under the BK21 project, and by Korean Ministry of Information and Communication under HNRC-ITRC program at Chung-Ang university supervised by IITA.