Accurate 3D Object Detection using Energy-Based Models

TL;DR

This paper introduces an energy-based model approach for 3D object detection that improves accuracy over existing methods by integrating a novel differentiable pooling operator into a state-of-the-art detector, demonstrating superior performance on KITTI.

Contribution

It presents a new EBM-based regression method with a differentiable pooling operator for 3D bounding boxes, enhancing 3D object detection accuracy.

Findings

Outperforms SA-SSD baseline on KITTI dataset

Achieves higher 3D detection metrics

Demonstrates the effectiveness of EBM in 3DOD

Abstract

Accurate 3D object detection (3DOD) is crucial for safe navigation of complex environments by autonomous robots. Regressing accurate 3D bounding boxes in cluttered environments based on sparse LiDAR data is however a highly challenging problem. We address this task by exploring recent advances in conditional energy-based models (EBMs) for probabilistic regression. While methods employing EBMs for regression have demonstrated impressive performance on 2D object detection in images, these techniques are not directly applicable to 3D bounding boxes. In this work, we therefore design a differentiable pooling operator for 3D bounding boxes, serving as the core module of our EBM network. We further integrate this general approach into the state-of-the-art 3D object detector SA-SSD. On the KITTI dataset, our proposed approach consistently outperforms the SA-SSD baseline across all 3DOD metrics, demonstrating the potential of EBM-based regression for highly accurate 3DOD. Code is available at https://github.com/fregu856/ebms_3dod.