Self-distillation and Uncertainty Boosting Self-supervised Monocular Depth Estimation
Hang Zhou (University of East Anglia),* Sarah Taylor (University of East Anglia), David Greenwood (University of East Anglia), Michal Mackiewicz (University of East Anglia)
The 33rd British Machine Vision Conference
Abstract
For self-supervised monocular depth estimation (SDE), recent works have introduced additional learning objectives, for example, semantic segmentation, into the training pipeline and have demonstrated improved performance. However, such multi-task learning frameworks require extra ground truth labels, neutralizing the biggest advantage of self-supervision. In this paper we propose SUB-Depth, a universal multi-task training framework, to overcome these limitations. Our main contribution is; we design an auxiliary self-distillation scheme and incorporate it into the standard SDE framework, to take advantage of multi-task learning without labeling cost. Then, instead of using a simple weighted sum of the multiple objectives, we employ generative task-dependent uncertainty to weight each task in our proposed training framework. We present extensive evaluations on KITTI to demonstrate the improvements achieved by training a range of existing networks using the proposed framework, and we achieve state-of-the-art performance on this task.Video
Citation
@inproceedings{Zhou_2022_BMVC,
author = {Hang Zhou and Sarah Taylor and David Greenwood and Michal Mackiewicz},
title = {Self-distillation and Uncertainty Boosting Self-supervised Monocular Depth Estimation},
booktitle = {33rd British Machine Vision Conference 2022, {BMVC} 2022, London, UK, November 21-24, 2022},
publisher = {{BMVA} Press},
year = {2022},
url = {https://bmvc2022.mpi-inf.mpg.de/0007.pdf}
}
Copyright © 2022
The British Machine Vision Conference is organised by