Study Of Face Feature Extraction In Driver Smart Monitor System

Driver smart monitor system monitors driving habitus by analyzing and understanding head sequences of driver in real time. A real-time algorithm for extracting facial features is presented in this paper. Main contents and outcome of the study is presented as follows.1. Face Features Extraction of Detecting Sequences. The facial organ region is localized based on horizontal and vertical gray-level projection, then extracting initiating parameter of mouth and eyes with morphology and pixels labeling methods. Last obtaining precise facial parameter by adjusting parameter based on symmetry of the face. The algorithm is simplified and useful.2. Face Features Extraction of Tracking Sequences. In view of actual application. A real-time algorithm based on mouth tracking for tracking faces and extracting facial features is presented in this paper. Meanwhile, the analysis of the symmetry of face guarantees precise tracking and extracting. Face violently mobile analysis is adopt to improve the real-time and robust of the algorithm.3. The algorithm was achieved by using VC++6.0. Experiment results show that features extraction algorithm can satiety real-tine and robust of driver smart monitor system and has high actual application value.