Class-balanced Grouping and Sampling for Point Cloud 3D Object Detection

TL;DR

This paper introduces a class-balanced grouping and sampling method for 3D object detection in point clouds, achieving state-of-the-art results in autonomous driving scenarios by addressing class imbalance and shape similarity issues.

Contribution

The paper proposes a novel class-balanced sampling, augmentation strategy, and a balanced grouping head to improve 3D detection performance in autonomous driving.

Findings

Outperforms PointPillars baseline significantly

Achieves state-of-the-art detection on nuScenes dataset

Effectively handles class imbalance and shape similarity

Abstract

This report presents our method which wins the nuScenes3D Detection Challenge [17] held in Workshop on Autonomous Driving(WAD, CVPR 2019). Generally, we utilize sparse 3D convolution to extract rich semantic features, which are then fed into a class-balanced multi-head network to perform 3D object detection. To handle the severe class imbalance problem inherent in the autonomous driving scenarios, we design a class-balanced sampling and augmentation strategy to generate a more balanced data distribution. Furthermore, we propose a balanced group-ing head to boost the performance for the categories withsimilar shapes. Based on the Challenge results, our methodoutperforms the PointPillars [14] baseline by a large mar-gin across all metrics, achieving state-of-the-art detection performance on the nuScenes dataset. Code will be released at CBGS.