Yumer and Mitra use 3D convolutional neural networks to predict semantic shape deformations. In particular, they consider shoes, cars, chairs and airplanes and deformations such as "comfy" for shoes or "sporty" for cars. The used network architecture corresponds to an encoder decoder scheme as illustrated in Figure 1. Results are shown in Figure 2.
Figure 1 (click to enlarge): The proposed network architectures to predict semantic shape deformations.
Figure 2 (click to enlarge): Qualitative results.
What is your opinion on the summarized work? Or do you know related work that is of interest? Let me know your thoughts in the comments below: