3D Human Face Reconstruction Based On RGB-D Image

Computer vision research how the computer”see the world”, and understand what itsee, With the development of computer technology, what computer can see is not only limitedto the color information, but also Depth information.3D technology developed a lot afterMicrosoft released its Kinect in2010.As one of the most important feature of human being, face always draw a lot attentionof computer vision researchers. Research on face mainly focus on recognition. Also2D facerecognition have made many breakthrough, but due to the limitations of the information of2D face, there are many problems remained to be solved. Therefore,3D face recognitionbecome an active field recently. In addition, we can build HCI system based on3D face.So3D face reconstruction is a problem urgent to be solved. In this paper, we makefull use of the exiting algorithm of computer version, build a3D face reconstruction systembased on RGB-D image. The algorithm used include: Viola-Jones face detect, Mean-shiftface tracking, image segmentation, ICP point cloud registration.