End-to-end deep stereo regression architecture

https://arxiv.org/abs/1703.04309

코드 : GC-Net, pythorch, TF, TF

- A deep learning architecture for regressing disparity from a rectified pair of stereo images.
- Leverage knowledge of the problem’s geometry to form a cost volume using deep feature representations.
- Learn to incorporate contextual information using 3-D convolutions over this volume.
- Disparity values regressed from the cost volume using a differentiable soft argmin operation, which allows to train end-to-end to sub-pixel accuracy without any additional post-processing or regularization.

참고 : 2018, Pyramid Stereo Matching Network, [code]