On the other hand, the parametric models like SMPL and SMPL-X utilize thousands of meshes to represent human body and can reconstruct a more detailed human body. However, it is impractical to support the reasoning about occlusion and collision and tell the somatotype of the estimated body only with the position of joints. For example, the skeleton is a simple and effective representation for the 3D pose estimation.
The reason is lacking a proper body model with efficient body part representation in both 2D and 3D.
READ FULL TEXT VIEW PDFĬurrent representations for human body succeed in developing 3D pose estimation, but still remain a great number of ambiguities. Our methods achieve the state-of-the-art results on Human3.6M and LSP datasetįor 3D pose estimation and part segmentation. Performance and efficiency at the same time. A small number of faces is chosen for achieving good Relationship between forward time, performance and number of faces in body
Which uses ellipsoids to indicate each body part. To further improve theĮfficiency of the task, we propose a light-weight body model called EllipBody, It enhances the performance inīoth learning-based and optimization-based methods. Which model occlusion between parts explicitly. To better utilize 3D informationĬontained in part segmentation, we propose a part-level differentiable renderer The shape of parts, as indicated in Figure 1.
Location of each part but also contains 3D information through occlusions from We find that part segmentation is a veryĮfficient 2D annotation in 3D human body recovery. Current works are tackled due to the lack of 3DĪnnotations for whole body shapes. Human pose and shape recovery is an important task in computer vision and