Learning 3D Shapes as Multi-Layered Height-Maps Using 2D Convolutional Networks
Abstract
We present a novel global representation of 3D shapes, suitable for the application of 2D CNNs. We represent 3D shapes as multi-layered height maps (MLH) where at each grid location, we store multiple instances of height maps, thereby representing 3D shape detail that is hidden behind several layers of occlusion. We provide a novel view merging method for combining view dependent information (Eg. MLH descriptors) from multiple views. Because of the ability of using 2D CNNs our method is highly memory efficient in terms of input resolution compared to the voxel based input. Together with MLH descriptors, and our multi view merging we achieve the state-of-the-art result in classification on ModelNet dataset.
Cite
Text
Sarkar et al. "Learning 3D Shapes as Multi-Layered Height-Maps Using 2D Convolutional Networks." Proceedings of the European Conference on Computer Vision (ECCV), 2018. doi:10.1007/978-3-030-01270-0_5Markdown
[Sarkar et al. "Learning 3D Shapes as Multi-Layered Height-Maps Using 2D Convolutional Networks." Proceedings of the European Conference on Computer Vision (ECCV), 2018.](https://mlanthology.org/eccv/2018/sarkar2018eccv-learning/) doi:10.1007/978-3-030-01270-0_5BibTeX
@inproceedings{sarkar2018eccv-learning,
title = {{Learning 3D Shapes as Multi-Layered Height-Maps Using 2D Convolutional Networks}},
author = {Sarkar, Kripasindhu and Hampiholi, Basavaraj and Varanasi, Kiran and Stricker, Didier},
booktitle = {Proceedings of the European Conference on Computer Vision (ECCV)},
year = {2018},
doi = {10.1007/978-3-030-01270-0_5},
url = {https://mlanthology.org/eccv/2018/sarkar2018eccv-learning/}
}