HVNet: Hybrid Voxel Network for LiDAR Based 3D Object Detection

TL;DR

HVNet introduces a hybrid voxel approach for LiDAR 3D object detection, combining multi-scale features and attention mechanisms to improve accuracy and speed for autonomous driving applications.

Contribution

The paper proposes a novel hybrid voxel network that fuses multi-scale voxel features and employs attentive encoding, achieving state-of-the-art results efficiently.

Findings

Achieves the highest mAP on KITTI benchmark among existing methods.

Operates at a real-time speed of 31Hz.

Outperforms plain VFE with an attentive voxel feature encoding.

Abstract

We present Hybrid Voxel Network (HVNet), a novel one-stage unified network for point cloud based 3D object detection for autonomous driving. Recent studies show that 2D voxelization with per voxel PointNet style feature extractor leads to accurate and efficient detector for large 3D scenes. Since the size of the feature map determines the computation and memory cost, the size of the voxel becomes a parameter that is hard to balance. A smaller voxel size gives a better performance, especially for small objects, but a longer inference time. A larger voxel can cover the same area with a smaller feature map, but fails to capture intricate features and accurate location for smaller objects. We present a Hybrid Voxel network that solves this problem by fusing voxel feature encoder (VFE) of different scales at point-wise level and project into multiple pseudo-image feature maps. We further propose an attentive voxel feature encoding that outperforms plain VFE and a feature fusion pyramid network to aggregate multi-scale information at feature map level. Experiments on the KITTI benchmark show that a single HVNet achieves the best mAP among all existing methods with a real time inference speed of 31Hz.