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

被引：91

作者：

Facil, Jose M. ^{[1
]}

Ummenhofer, Benjamin ^{[2
,3
]}

Zhou, Huizhong ^{[2
]}

Montesano, Luis ^{[1
,4
]}

Brox, Thomas ^{[2
]}

Civera, Javier ^{[1
]}

机构：

[1] Univ Zaragoza, Zaragoza, Spain

[2] Univ Freiburg, Freiburg, Germany

[3] Intel Labs, Hillsboro, OR USA

[4] Bitbrain, Zaragoza, Spain

来源：

2019 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR 2019) | 2019年

基金：

欧盟地平线“2020”;

关键词：

D O I：

10.1109/CVPR.2019.01210

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

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.

引用

页码：11818 / 11827

页数：10