MAT-Net: Medial Axis Transform Network for 3D Object Recognition

Jianwei Hu, Bin Wang, Lihui Qian, Yiling Pan, Xiaohu Guo, Lingjie Liu, Wenping Wang

IJCAI 2019 pp. 774-781

doi:10.24963/IJCAI.2019/109 /ijcai/2019/hu2019ijcai-mat/

Abstract

3D deep learning performance depends on object representation and local feature extraction. In this work, we present MAT-Net, a neural network which captures local and global features from the Medial Axis Transform (MAT). Different from K-Nearest-Neighbor method which extracts local features by a fixed number of neighbors, our MAT-Net exploits effective modules Group-MAT and Edge-Net to process topological structure. Experimental results illustrate that MAT-Net demonstrates competitive or better performance on 3D shape recognition than state-of-the-art methods, and prove that MAT representation has excellent capacity in 3D deep learning, even in the case of low resolution.

PDF IJCAI Semantic Scholar

Cite

Text

Hu et al. "MAT-Net: Medial Axis Transform Network for 3D Object Recognition." International Joint Conference on Artificial Intelligence, 2019. doi:10.24963/IJCAI.2019/109

Markdown

[Hu et al. "MAT-Net: Medial Axis Transform Network for 3D Object Recognition." International Joint Conference on Artificial Intelligence, 2019.](https://mlanthology.org/ijcai/2019/hu2019ijcai-mat/) doi:10.24963/IJCAI.2019/109

BibTeX

@inproceedings{hu2019ijcai-mat,
  title     = {{MAT-Net: Medial Axis Transform Network for 3D Object Recognition}},
  author    = {Hu, Jianwei and Wang, Bin and Qian, Lihui and Pan, Yiling and Guo, Xiaohu and Liu, Lingjie and Wang, Wenping},
  booktitle = {International Joint Conference on Artificial Intelligence},
  year      = {2019},
  pages     = {774-781},
  doi       = {10.24963/IJCAI.2019/109},
  url       = {https://mlanthology.org/ijcai/2019/hu2019ijcai-mat/}
}