Differentiable Convex Polyhedra Optimization from Multi-View Images

Abstract

This paper presents a novel approach for the differentiable rendering of convex polyhedra, addressing the limitations of recent methods that rely on implicit field supervision. Our technique introduces a strategy that combines non-differentiable computation of hyperplane intersection through duality transform with differentiable optimization for vertex positioning with three-plane intersection, enabling gradient-based optimization without the need for 3D implicit fields. This allows for efficient shape representation across a range of applications, from shape parsing to compact mesh reconstruction. This work not only overcomes the challenges of previous approaches but also sets a new standard for representing shapes with convex polyhedra.

Cite

Text

Ren et al. "Differentiable Convex Polyhedra Optimization from Multi-View Images." Proceedings of the European Conference on Computer Vision (ECCV), 2024. doi:10.1007/978-3-031-72673-6_14

Markdown

[Ren et al. "Differentiable Convex Polyhedra Optimization from Multi-View Images." Proceedings of the European Conference on Computer Vision (ECCV), 2024.](https://mlanthology.org/eccv/2024/ren2024eccv-differentiable/) doi:10.1007/978-3-031-72673-6_14

BibTeX

@inproceedings{ren2024eccv-differentiable,
  title     = {{Differentiable Convex Polyhedra Optimization from Multi-View Images}},
  author    = {Ren, Daxuan and Mei, Haiyi and Shi, Hezi and Zheng, Jianmin and Cai, Jianfei and Yang, Lei},
  booktitle = {Proceedings of the European Conference on Computer Vision (ECCV)},
  year      = {2024},
  doi       = {10.1007/978-3-031-72673-6_14},
  url       = {https://mlanthology.org/eccv/2024/ren2024eccv-differentiable/}
}