A State-of-the-Art Computer Vision Adopting Non-Euclidean Deep-Learning Models

Figure 11

(a) Shape analysis with different views; (b) the multiview model architecture takes views as 2D input to analyze 3D objects.