Casting Geometric Constraints in Semantic Segmentation as Semi-Supervised Learning

Open in new window