Rethinking the Encoding and Annotating of 3D Bounding Box: Corner-Aware 3D Object Detection from Point Clouds

TL;DR

This paper introduces corner-aware regression for 3D object detection from point clouds, replacing unstable center-based predictions with geometrically informative corners, leading to improved accuracy and weakly supervised learning.

Contribution

It proposes a novel corner-aligned regression method that enhances 3D detection accuracy and enables weak supervision using only corner annotations.

Findings

Improves detection performance by 3.5% AP on KITTI.

Achieves 83% of fully supervised accuracy with only corner clicks.

Demonstrates effectiveness of corner-aware regression in real-world data.

Abstract

Center-aligned regression remains dominant in LiDAR-based 3D object detection, yet it suffers from fundamental instability: object centers often fall in sparse or empty regions of the bird's-eye-view (BEV) due to the front-surface-biased nature of LiDAR point clouds, leading to noisy and inaccurate bounding box predictions. To circumvent this limitation, we revisit bounding box representation and propose corner-aligned regression, which shifts the prediction target from unstable centers to geometrically informative corners that reside in dense, observable regions. Leveraging the inherent geometric constraints among corners and image 2D boxes, partial parameters of 3D bounding boxes can be recovered from corner annotations, enabling a weakly supervised paradigm without requiring complete 3D labels. We design a simple yet effective corner-aware detection head that can be plugged into existing detectors. Experiments on KITTI show our method improves performance by 3.5% AP over center-based baseline, and achieves 83% of fully supervised accuracy using only BEV corner clicks, demonstrating the effectiveness of our corner-aware regression strategy.