3D human body inpainting using intrinsic statistical shape models
This paper presents an algorithm for inpainting 3D human body geometry. We built a Statistical Shape Model (SSM) based on the intrinsic Laplacian Coordinates (LC) to tackle the pose variation of human bodies, which is a common challenging issue in 3D human body modeling and shape analysis. We evaluate our inpainting algorithm using synthesized damages on complete human body scans, and with the groundtruth we measure the normalized reconstruction error quantitatively. Our experiments demonstrate the effectiveness of this LC-based SSM in human body inpainting.
Publication Source (Journal or Book title)
14th International Conference on Computer Science and Education, ICCSE 2019
Korban, M., Li, X., Miao, K., & Zhu, Y. (2019). 3D human body inpainting using intrinsic statistical shape models. 14th International Conference on Computer Science and Education, ICCSE 2019, 1105-1110. https://doi.org/10.1109/ICCSE.2019.8845520