Object DGCNN: 3D Object Detection using Dynamic Graphs

TL;DR

This paper introduces Object DGCNN, a novel 3D object detection method on point clouds that eliminates post-processing steps and employs a set-to-set distillation approach, achieving state-of-the-art results in autonomous driving benchmarks.

Contribution

It generalizes the DGCNN framework for 3D detection, removing the need for non-maximum suppression and introducing a permutation-invariant distillation method.

Findings

Achieves state-of-the-art performance on autonomous driving benchmarks.

Eliminates post-processing steps like non-maximum suppression.

Introduces a set-to-set distillation approach for 3D detection.

Abstract

3D object detection often involves complicated training and testing pipelines, which require substantial domain knowledge about individual datasets. Inspired by recent non-maximum suppression-free 2D object detection models, we propose a 3D object detection architecture on point clouds. Our method models 3D object detection as message passing on a dynamic graph, generalizing the DGCNN framework to predict a set of objects. In our construction, we remove the necessity of post-processing via object confidence aggregation or non-maximum suppression. To facilitate object detection from sparse point clouds, we also propose a set-to-set distillation approach customized to 3D detection. This approach aligns the outputs of the teacher model and the student model in a permutation-invariant fashion, significantly simplifying knowledge distillation for the 3D detection task. Our method achieves state-of-the-art performance on autonomous driving benchmarks. We also provide abundant analysis of the detection model and distillation framework.