O-CNN: Octree-based Convolutional Neural Networks for 3D Shape Analysis

TL;DR

O-CNN introduces an octree-based 3D CNN that efficiently analyzes high-resolution 3D shapes by focusing computations on surface-occupied octants, enabling effective shape classification, retrieval, and segmentation.

Contribution

The paper presents a novel octree data structure and GPU implementation for 3D CNNs, improving efficiency and scalability for high-resolution shape analysis.

Findings

Outperforms existing 3D CNN methods in efficiency and accuracy.

Supports various CNN architectures for 3D shape tasks.

Enables high-resolution 3D shape analysis with manageable computational costs.

Abstract

We present O-CNN, an Octree-based Convolutional Neural Network (CNN) for 3D shape analysis. Built upon the octree representation of 3D shapes, our method takes the average normal vectors of a 3D model sampled in the finest leaf octants as input and performs 3D CNN operations on the octants occupied by the 3D shape surface. We design a novel octree data structure to efficiently store the octant information and CNN features into the graphics memory and execute the entire O-CNN training and evaluation on the GPU. O-CNN supports various CNN structures and works for 3D shapes in different representations. By restraining the computations on the octants occupied by 3D surfaces, the memory and computational costs of the O-CNN grow quadratically as the depth of the octree increases, which makes the 3D CNN feasible for high-resolution 3D models. We compare the performance of the O-CNN with other existing 3D CNN solutions and demonstrate the efficiency and efficacy of O-CNN in three shape analysis tasks, including object classification, shape retrieval, and shape segmentation.