Unsupervised Scale-consistent Depth Learning from Video
| Autor: | Bian, Jia-Wang, Zhan, Huangying, Wang, Naiyan, Li, Zhichao, Zhang, Le, Shen, Chunhua, Cheng, Ming-Ming, Reid, Ian |
|---|---|
| Rok vydání: | 2021 |
| Předmět: | |
| Zdroj: | International Journal of Computer Vision, 2021 |
| Druh dokumentu: | Working Paper |
| DOI: | 10.1007/s11263-021-01484-6 |
| Popis: | We propose a monocular depth estimator SC-Depth, which requires only unlabelled videos for training and enables the scale-consistent prediction at inference time. Our contributions include: (i) we propose a geometry consistency loss, which penalizes the inconsistency of predicted depths between adjacent views; (ii) we propose a self-discovered mask to automatically localize moving objects that violate the underlying static scene assumption and cause noisy signals during training; (iii) we demonstrate the efficacy of each component with a detailed ablation study and show high-quality depth estimation results in both KITTI and NYUv2 datasets. Moreover, thanks to the capability of scale-consistent prediction, we show that our monocular-trained deep networks are readily integrated into the ORB-SLAM2 system for more robust and accurate tracking. The proposed hybrid Pseudo-RGBD SLAM shows compelling results in KITTI, and it generalizes well to the KAIST dataset without additional training. Finally, we provide several demos for qualitative evaluation. Comment: Accept to IJCV. The source code is available at https://github.com/JiawangBian/SC-SfMLearner-Release |
| Databáze: | arXiv |
| Externí odkaz: |
|