We present a novel convolutional neural network architecture for photometric stereo (Woodham, 1980), a problem of recovering 3D object surface normals from multiple images observed under varying illuminations.
In this paper we evaluate plane orientation priors derived from stereo matching at a coarser resolution and show that such priors can yield significant performance gains for difficult weakly-textured scenes.
This unified framework benefits all four tasks - stereo, optical flow, visual odometry and motion segmentation leading to overall higher accuracy and efficiency.
The local expansion moves extend traditional expansion moves by two ways: localization and spatial propagation.
We then present our method as generalization of SSP, which is further shown to generalize several state-of-the-art techniques for higher-order and pairwise non-submodular functions [Ayed13, Gorelick14, Tang14].
We present an accurate and efficient stereo matching method using locally shared labels, a new labeling scheme that enables spatial propagation in MRF inference using graph cuts.