The demo of the Paper “Occlusion-Aware Unsupervised Learning of Depth and Camera Pose With Feature Map Warping”.
The video of demo on bilibili.
If this paper is accepted, we will consider opening source code.
You can download our precomputed disparity predictions from the following links:
| Training modality | .npy filesize |
Eigen disparities |
|---|---|---|
| +F | 326MB | Download |
| +OM | 326MB | Download |
| +OM+F | 326MB | Download |