by Marcela Carvalho, Bertrand Le Saux, Pauline Trouve-Peloux, Andres Almansa, Frederic Champagnat
Abstract:
Depth estimation from a single monocular image has reached great performances thanks to recent works based on deep networks. However, as various choices of losses, architectures and experimental conditions are proposed in the literature, it is difficult to establish their respective influence on the performances. In this paper we propose an in-depth study of various losses and experimental conditions for depth regression, on NYUv2 dataset. From this study we propose a new network for depth estimation combining an encoder-decoder architecture with an adversarial loss. This network reaches top scores in the competitive evaluation of NUYv2 dataset while being simpler to train in a single phase.
Reference:
On Regression Losses for Deep Depth Estimation (Marcela Carvalho, Bertrand Le Saux, Pauline Trouve-Peloux, Andres Almansa, Frederic Champagnat), In (ICIP) IEEE International Conference on Image Processing, IEEE, 2018.
Bibtex Entry:
@inproceedings{Carvalho2018-ICIP,
Abstract = {Depth estimation from a single monocular image has reached great performances thanks to recent works based on deep networks. However, as various choices of losses, architectures and experimental conditions are proposed in the literature, it is difficult to establish their respective influence on the performances. In this paper we propose an in-depth study of various losses and experimental conditions for depth regression, on NYUv2 dataset. From this study we propose a new network for depth estimation combining an encoder-decoder architecture with an adversarial loss. This network reaches top scores in the competitive evaluation of NUYv2 dataset while being simpler to train in a single phase.},
Address = {Athens},
Author = {Carvalho, Marcela and Saux, Bertrand Le and Trouve-Peloux, Pauline and Almansa, Andres and Champagnat, Frederic},
Booktitle = {(ICIP) IEEE International Conference on Image Processing},
Doi = {10.1109/ICIP.2018.8451312},
Isbn = {978-1-4799-7061-2},
Month = {oct},
Pages = {2915--2919},
Publisher = {IEEE},
Title = {{On Regression Losses for Deep Depth Estimation}},
Url = {https://github.com/marcelampc/d3net_depth_estimation},
Year = {2018},
Bdsk-Url-1 = {https://github.com/marcelampc/d3net%7B%5C_%7Ddepth%7B%5C_%7Destimation},
Bdsk-Url-2 = {https://doi.org/10.1109/ICIP.2018.8451312}}