THREE-DIMENSIONAL RECONSTRUCTION METHOD AND SYSTEM FOR JOINT OPTIMIZATION OF HUMAN BODY POSTURE MODEL AND APPEARANCE MODEL

摘要

A three-dimensional reconstruction method and system for the joint optimization of a human body posture model and an appearance model. The method comprises the following steps: capturing a depth map of a human body to obtain a single depth image (S101); converting the single depth image into a three-dimensional point cloud, and obtaining matching point pairs between the three-dimensional point cloud and vertexes of a reconstruction model and matching point pairs between the three-dimensional point cloud and vertexes of a parameterized human body model (S102); building an energy function according to the matching point pairs, and jointly solving a non-rigid motion position transformation parameter of each vertex on the reconstruction model and a parameterized human body posture model parameter (S103); solving the energy function, and performing alignment on the reconstruction model and the three-dimensional point cloud according to the solution result (S104); and updating and complementing the aligned model by means of the depth map to perform real-time human body dynamic three-dimensional reconstruction (S105). The method can effectively improve the timeliness, robustness and accuracy of reconstruction, features high expandability, and is simple and easy to implement.