ML Anthology

CAM-Convs: Camera-Aware Multi-Scale Convolutions for Single-View Depth

Jose M. Facil, Benjamin Ummenhofer, Huizhong Zhou, Luis Montesano, Thomas Brox, Javier Civera
CVPR 2019
doi:10.1109/CVPR.2019.01210 /cvpr/2019/facil2019cvpr-camconvs/

Abstract

Single-view depth estimation suffers from the problem that a network trained on images from one camera does not generalize to images taken with a different camera model. Thus, changing the camera model requires collecting an entirely new training dataset. In this work, we propose a new type of convolution that can take the camera parameters into account, thus allowing neural networks to learn calibration-aware patterns. Experiments confirm that this improves the generalization capabilities of depth prediction networks considerably, and clearly outperforms the state of the art when the train and test images are acquired with different cameras.

PDF CVPR Semantic Scholar

Cite

Text

Facil et al. "CAM-Convs: Camera-Aware Multi-Scale Convolutions for Single-View Depth." Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2019. doi:10.1109/CVPR.2019.01210

Markdown

[Facil et al. "CAM-Convs: Camera-Aware Multi-Scale Convolutions for Single-View Depth." Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2019.](https://mlanthology.org/cvpr/2019/facil2019cvpr-camconvs/) doi:10.1109/CVPR.2019.01210

BibTeX

@inproceedings{facil2019cvpr-camconvs,
  title     = {{CAM-Convs: Camera-Aware Multi-Scale Convolutions for Single-View Depth}},
  author    = {Facil, Jose M. and Ummenhofer, Benjamin and Zhou, Huizhong and Montesano, Luis and Brox, Thomas and Civera, Javier},
  booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
  year      = {2019},
  doi       = {10.1109/CVPR.2019.01210},
  url       = {https://mlanthology.org/cvpr/2019/facil2019cvpr-camconvs/}
}