Accurate 3D Object Detection using Energy-Based Models

8 Dec 2020  ·  Fredrik K. Gustafsson, Martin Danelljan, Thomas B. Schön ·

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.

PDF Abstract

Datasets


Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
3D Object Detection KITTI Cars Easy SA-SSD+EBM AP 91.05% # 3
3D Object Detection KITTI Cars Easy val SA-SSD+EBM AP 95.45 # 1
3D Object Detection KITTI Cars Hard SA-SSD+EBM AP 72.78% # 13
3D Object Detection KITTI Cars Hard val SA-SSD+EBM AP 82.23 # 3
3D Object Detection KITTI Cars Moderate SA-SSD+EBM AP 80.12% # 13
3D Object Detection KITTI Cars Moderate val SA-SSD+EBM AP 86.83 # 1

Methods