3D Face Detection

Face feature is a very common and very complex visual pattern could reflect the face of visual information in people's lives, play an important role and significance of human face processing and analysis of information in social security, public video surveillance, entrance test, online video tracking, human-computer interaction and other areas with wide application, so long as the field of pattern recognition and computer vision continued research focus.In recent years, due to facial feature extraction in face processing and analysis of the position, facial feature extraction has attracted the attention of more and more researchers, and gradually developed into a relatively independent research, the research methods and maturity also means the emergence of a variety of practical face detection algorithm. The 3D visual appearance of the external environment can be a very good customer service impact factor for face recognition, face recognition can solve encountered light, gesture, expression and so on.Face recognition feature extraction is an important part of the study, to effectively extract the facial feature recognition technology is the key to be able to face the complex graphics in a large number of data extracted features and classification extraction.In this paper pattern recognition theory and computer data in 3D facial features, based on the CAS Institute of Automation 3D face database CASIA, focus on the 3D face detection system in the various component modules in-depth research. Its main tasks are:1, Achieved 3D information on the point cloud data surface fitting, and rotate 3D spatial information, cutting, the profile separation.2, Achieved the rapid 3D nasal tip positioning algorithm, and use as the base area, on the face of different 3D data registration.3, Improved a 3D face data dimension reduction methods, making the resulting 3D contour face graph more accurate depth, demonstrated changes in the gray area more evenly. Experimental results show that the algorithm is robust, can be effectively used in different sizes, different expressions, different situation gesture.4, Experiment, by comparing the properties of discrete points on the feature extraction methods and properties of surface features based on surface feature extraction method, and the initial exploration and attempt to propose a correlation based on surface characteristics of the feature extraction method.Overall, this 3D information based on 3D human face data pre-processing and 3D human face feature extraction direction do in-depth study, and obtained some preliminary results, I hope this work can be made for further research and development to help .