Research on the Application of Animation Character Motion Capture Technology Based on Deep Learning in Micro-Film Production

Firstly, based on the two basic concepts of motion capture and cluster rendering, the necessity of building a 3D character dynamic rendering engine is proposed. The main functional modules, key technologies and programs of the software are introduced in detail. The research needs only a small amount of motion capture information to obtain skeleton motion information for different types of human feature modeling. The algorithm first uses motion capture data to train the motion data autoencoder, and then obtains the bone motion data of various human role models by iterative constraint solving in the hidden variable space generated by the autoencoder. The results show that the proposed algorithm has good performance. The model can generate different types of skeleton actions efficiently, which provides a new idea for further expanding the skeleton action database. It can reduce the cost of animation production for film and television, animation and games. Compared with the previous 3D animation production process, the 3D character animation creation machine designed in this paper has a great innovation in concept, technology and technology.