V-DETR: DETR with Vertex Relative Position Encoding for 3D Object Detection

TL;DR

V-DETR introduces a novel 3D vertex relative position encoding to improve 3D object detection in point clouds, significantly outperforming previous methods on benchmark datasets.

Contribution

The paper proposes a new 3D vertex relative position encoding method that enhances the DETR framework for better locality and accuracy in 3D object detection.

Findings

Achieves 77.8% AP25 on ScanNetV2, surpassing previous methods.

Sets a new record on ScanNetV2 and SUN RGB-D datasets.

Demonstrates significant performance improvements with the proposed encoding.

Abstract

We introduce a highly performant 3D object detector for point clouds using the DETR framework. The prior attempts all end up with suboptimal results because they fail to learn accurate inductive biases from the limited scale of training data. In particular, the queries often attend to points that are far away from the target objects, violating the locality principle in object detection. To address the limitation, we introduce a novel 3D Vertex Relative Position Encoding (3DV-RPE) method which computes position encoding for each point based on its relative position to the 3D boxes predicted by the queries in each decoder layer, thus providing clear information to guide the model to focus on points near the objects, in accordance with the principle of locality. In addition, we systematically improve the pipeline from various aspects such as data normalization based on our understanding of the task. We show exceptional results on the challenging ScanNetV2 benchmark, achieving significant improvements over the previous 3DETR in $\rm{AP}_{25}$/$\rm{AP}_{50}$ from 65.0\%/47.0\% to 77.8\%/66.0\%, respectively. In addition, our method sets a new record on ScanNetV2 and SUN RGB-D datasets.Code will be released at http://github.com/yichaoshen-MS/V-DETR.