3D ShapeNets: A Deep Representation for Volumetric Shapes

TL;DR

This paper introduces 3D ShapeNets, a deep learning model that effectively represents and recognizes 3D shapes from CAD data, enabling improved object recognition and shape completion from depth maps.

Contribution

It presents a novel deep probabilistic model for 3D shape representation that learns hierarchical features and supports joint recognition and shape reconstruction.

Findings

Outperforms previous methods in shape recognition accuracy

Enables effective shape completion from partial depth data

Supports active view planning for better recognition

Abstract

3D shape is a crucial but heavily underutilized cue in today's computer vision systems, mostly due to the lack of a good generic shape representation. With the recent availability of inexpensive 2.5D depth sensors (e.g. Microsoft Kinect), it is becoming increasingly important to have a powerful 3D shape representation in the loop. Apart from category recognition, recovering full 3D shapes from view-based 2.5D depth maps is also a critical part of visual understanding. To this end, we propose to represent a geometric 3D shape as a probability distribution of binary variables on a 3D voxel grid, using a Convolutional Deep Belief Network. Our model, 3D ShapeNets, learns the distribution of complex 3D shapes across different object categories and arbitrary poses from raw CAD data, and discovers hierarchical compositional part representations automatically. It naturally supports joint object recognition and shape completion from 2.5D depth maps, and it enables active object recognition through view planning. To train our 3D deep learning model, we construct ModelNet -- a large-scale 3D CAD model dataset. Extensive experiments show that our 3D deep representation enables significant performance improvement over the-state-of-the-arts in a variety of tasks.