Pix2Vex Image-to-Geometry Reconstruction using a Smooth Differentiable Renderer
Published:
threefold contributions
- C^{infty} smooth differentiable renderer
- image2geometry NN: predict 3D coordintes of given topology
- due to its diff. nature, DR’s output is diff. from real world images
- introduce a new module to map to real image field
- reconstructive adversial network
- to discriminate the mapping module mentioned above
renderer
- view transformation: triangles -> proj. coor.
- resterization: triangles -> pixels
- keep a z-buffer
- …of prob. of visible triangles
- softmin-based function: differentiable!
reconstructor
- pre-defined topology
- conv-deconv-fc-sigmoid