MVPNet: Multi-View Point Regression Networks for 3D Object Reconstruction from A Single Image

Understand

In this paper, we address the problem of reconstructing an object's surface from a single image using generative networks.

• First, we represent a 3D surface with an aggregation of dense point clouds from multiple views.
• Each point cloud is embedded in a regular 2D grid aligned on an image plane of a viewpoint, making the point cloud convolution-favored and ordered so as to fit into deep network architectures.
• The point clouds can be easily triangulated by exploiting connectivities of the 2D grids to form mesh-based surfaces.