基于人体骨骼的稠密三维重建

In the past few decades, 3D reconstruction using consumer-grade RGB-D cameras has been extensively studied in the fields of computer vision, robotics, and computer graphics. With the rapid development of Augmented Reality (AR) and Virtual Reality (VR), there have been many applications that require the establishment of 3D human models, including human tracking and monitoring and biometrics. Among them, the established 3D human body model should not only include the specific static geometry of the human body, hair, and clothing, but also should also include a coherent surface texture. This article is based on SegmentedFusion\'s work of segmenting the human body based on the skeleton information in the reconstruction of the human body. The surface deformation node map is constructed on the surface of the segmented body part. The bone movement and surface deformation are used to track the reconstructed human body\'s pose transformation. A dense warping field based on surface deformation nodes of human bones is proposed. Through experiments, it is proved that the warped dense field of surface deformation nodes based on human bones proposed in this paper effectively solves the problem of tracking torsional motion during 3D reconstruction of human body, and refines the reconstruction of surface deformation.