TripleDNet: Exploring Depth Estimation with Self-Supervised Representation Learning
Ufuk Umut Senturk (Hacettepe University),* Arif Akar (Hacettepe University), Nazli Ikizler-Cinbis (Hacettepe University)
The 33rd British Machine Vision Conference
Abstract
We propose TripleDNet (Disentangled Distilled Depth Network), a multi-objective, distillation-based framework for purely self-supervised depth estimation. %Structure-from-motion-based depth prediction models utilize self-supervision while processing consecutive frames in a monocular depth estimation manner. Static world and illumination constancy assumptions do not hold and allow wrong signals to the training procedure, leading to poor performance. Masking out those parts hurts the integrity of the image structure. We add further objectives to structure-from-motion-based estimation to constrain the solution space and to allow feature space disentanglement within an efficient and simple architecture. In addition, we propose a knowledge distillation objective that supports depth estimation in terms of scene context and structure. Surprisingly, we also found that self-supervised image representation learning frameworks for model initialization outperform the supervised counterparts. Experimental results show that proposed models trained purely in a self-supervised fashion outperform the state-of-the-art models on the KITTI and Make3D datasets compared to models utilizing ground truth segmentation maps. Codes are available at https://github.com/ufukpage/TripleD.Video
Citation
@inproceedings{Senturk_2022_BMVC,
author = {Ufuk Umut Senturk and Arif Akar and Nazli Ikizler-Cinbis},
title = {TripleDNet: Exploring Depth Estimation with Self-Supervised Representation Learning},
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/0008.pdf}
}
Copyright © 2022
The British Machine Vision Conference is organised by