Multiscale Region-Based Convolutional Neural Networks for 3D Object Detection with LiDAR Sensors

TL;DR

This paper introduces a new 3D object detection system using LiDAR data that improves performance through multiscale feature fusion and refined data augmentation.

Contribution

The novel contribution is a refinement fusion network with cross-attention modules for 3D object detection in LiDAR data.

Findings

The proposed system achieved improved results across three categories on the KITTI dataset.

The multiscale feature fusion mechanism enhanced object detection performance.

Refined data augmentation strategies contributed to the system's effectiveness.

Abstract

LiDAR-based 3D object detection is essential for autonomous driving vehicles under poor lighting conditions. With LiDAR data, point cloud technologies have become increasingly important, as LiDAR sensors are largely cost down. However, the sparsity of point cloud poses a challenge for 3D object detection, requiring advancements in sparse convolutional networks. Given that the multiscale feature fusion mechanism can improve object detection performance using rich information across scale features, we added a refinement fusion network with cross-attention modules to existing 3D voxel-based object detection networks. We also employed a realistic strategy to refine existing point cloud data augmentation techniques to enable the trained detection networks to achieve substantially improved results. The experimental results demonstrate the effectiveness of our proposed detection system across three categories on the KITTI dataset. These enhancements address the limitations of current approaches and highlight the superior performance of the proposed system.

Figures