3D Object Detection with Pointformer
This work provides an incremental improvement in 3D object detection performance for researchers and practitioners working with point cloud data.
This paper introduces Pointformer, a Transformer-based backbone for 3D object detection from point clouds. It significantly improves state-of-the-art object detection models on both indoor and outdoor datasets by learning context-dependent region features and context-aware scene representations.
Feature learning for 3D object detection from point clouds is very challenging due to the irregularity of 3D point cloud data. In this paper, we propose Pointformer, a Transformer backbone designed for 3D point clouds to learn features effectively. Specifically, a Local Transformer module is employed to model interactions among points in a local region, which learns context-dependent region features at an object level. A Global Transformer is designed to learn context-aware representations at the scene level. To further capture the dependencies among multi-scale representations, we propose Local-Global Transformer to integrate local features with global features from higher resolution. In addition, we introduce an efficient coordinate refinement module to shift down-sampled points closer to object centroids, which improves object proposal generation. We use Pointformer as the backbone for state-of-the-art object detection models and demonstrate significant improvements over original models on both indoor and outdoor datasets.