Semi-Supervised Learning of Optical Flow by Flow Supervisor
About
A training pipeline for optical flow CNNs consists of a pretraining stage on a synthetic dataset followed by a fine tuning stage on a target dataset. However, obtaining ground truth flows from a target video requires a tremendous effort. This paper proposes a practical fine tuning method to adapt a pretrained model to a target dataset without ground truth flows, which has not been explored extensively. Specifically, we propose a flow supervisor for self-supervision, which consists of parameter separation and a student output connection. This design is aimed at stable convergence and better accuracy over conventional self-supervision methods which are unstable on the fine tuning task. Experimental results show the effectiveness of our method compared to different self-supervision methods for semi-supervised learning. In addition, we achieve meaningful improvements over state-of-the-art optical flow models on Sintel and KITTI benchmarks by exploiting additional unlabeled datasets. Code is available at https://github.com/iwbn/flow-supervisor.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Optical Flow | Sintel (test) | AEPE (Final)2.79 | 120 | |
| Optical Flow | Sintel Final (train) | EPE2.46 | 92 | |
| Optical Flow | Sintel Clean (train) | EPE1.3 | 85 | |
| Optical Flow | KITTI (train) | Fl-all0.159 | 63 | |
| Optical Flow | KITTI (test) | -- | 28 | |
| Optical Flow Estimation | KITTI (test) | F1-all4.85 | 20 |