Expression dynamic capture and 3D animation generation method based on deep learning

With the rapid development of computer technology and artificial intelligence today, human-computer interaction based solely on computer software operations is far from being able to meet human requirements for computer use. People are looking forward to a more convenient and fast human-computer interface. This paper studies the expression dynamic capture and 3D animation generation methods based on deep learning. For facial expression dynamic capture, this paper proposes a facial feature extraction algorithm based on deep learning and uses SVM technology for feature classification. For 3D animation, C++ and OpenGL are used for rendering simulation. The experimental results show that the face detection algorithm proposed in this paper has good performance in both accuracy and speed. It can realize real-time detection of face regions in video images.