Abstract:When learning to sketch, beginners start with simple and flexible shapes, and then gradually strive for more complex and accurate ones in the subsequent training sessions. In this paper, we design a "shape curriculum" for learning continuous Signed Distance Function (SDF) on shapes, namely Curriculum DeepSDF. Inspired by how humans learn, Curriculum DeepSDF organizes the learning task in ascending order of difficulty according to the following two criteria: surface accuracy and sample difficulty. The former considers stringency in supervising with ground truth, while the latter regards the weights of hard training samples near complex geometry and fine structure. More specifically, Curriculum DeepSDF learns to reconstruct coarse shapes at first, and then gradually increases the accuracy and focuses more on complex local details. Experimental results show that a carefully-designed curriculum leads to significantly better shape reconstructions with the same training data, training epochs and network architecture as DeepSDF. We believe that the application of shape curricula can benefit the training process of a wide variety of 3D shape representation learning methods.
| Comments: | ECCV 2020 |
| Subjects: | Computer Vision and Pattern Recognition (cs.CV) |
| Cite as: | arXiv:2003.08593 [cs.CV] |
| (or arXiv:2003.08593v3 [cs.CV] for this version) | |
| https://doi.org/10.48550/arXiv.2003.08593 arXiv-issued DOI via DataCite |
Submission history
From: Yueqi Duan [view email]
[v1]
Thu, 19 Mar 2020 06:40:59 UTC (790 KB)
[v2]
Tue, 14 Jul 2020 19:23:49 UTC (791 KB)
[v3]
Thu, 16 Jul 2020 21:18:48 UTC (791 KB)