Multi-View 3D Object Detection Network for Autonomous Driving

TL;DR

This paper introduces MV3D, a multi-view sensory fusion network for high-accuracy 3D object detection in autonomous driving, combining LIDAR and camera data to outperform existing methods.

Contribution

The paper presents a novel multi-view fusion framework that efficiently encodes sparse point clouds and fuses multi-view features for improved 3D detection accuracy.

Findings

Outperforms state-of-the-art by 25-30% AP on KITTI benchmark

Achieves 10.3% higher AP in 2D detection on hard data

Effective multi-view fusion improves 3D localization and detection

Abstract

This paper aims at high-accuracy 3D object detection in autonomous driving scenario. We propose Multi-View 3D networks (MV3D), a sensory-fusion framework that takes both LIDAR point cloud and RGB images as input and predicts oriented 3D bounding boxes. We encode the sparse 3D point cloud with a compact multi-view representation. The network is composed of two subnetworks: one for 3D object proposal generation and another for multi-view feature fusion. The proposal network generates 3D candidate boxes efficiently from the bird's eye view representation of 3D point cloud. We design a deep fusion scheme to combine region-wise features from multiple views and enable interactions between intermediate layers of different paths. Experiments on the challenging KITTI benchmark show that our approach outperforms the state-of-the-art by around 25% and 30% AP on the tasks of 3D localization and 3D detection. In addition, for 2D detection, our approach obtains 10.3% higher AP than the state-of-the-art on the hard data among the LIDAR-based methods.