WHAT |
A novel idea of incorporating 3D deformation of point cloud in a deep learned model, applied to 3D reconstruction |
WHEN |
January 2017-May 2017 |
WHO |
Andrey Kurenkov, Jingwei Ji, Animesh Garg, Viraj Mehta, JunYoung Gwak, Christopher Choy, Silvio Savarese |
WHERE |
Stanford Vision Lab |
WHY |
Deformations from similar objets based on shape retrieval is an intuitively reasonable approach to 3D reconstruction, and so this was a good idea. |
HOW |
I was not the one to have the original idea, but mostly the one to grind out and iterate a lot to get it working. See the Arxiv post and website for more details. |
LINKS |