Pillar-based Object Detection for Autonomous Driving
This work improves object detection for autonomous vehicles by eliminating anchor-related hyperparameters while enhancing accuracy.
The paper tackles 3D object detection for autonomous driving by proposing a pillar-based anchor-free framework that addresses data sparsity and anchor imbalance issues, achieving state-of-the-art performance with simplified implementation.
We present a simple and flexible object detection framework optimized for autonomous driving. Building on the observation that point clouds in this application are extremely sparse, we propose a practical pillar-based approach to fix the imbalance issue caused by anchors. In particular, our algorithm incorporates a cylindrical projection into multi-view feature learning, predicts bounding box parameters per pillar rather than per point or per anchor, and includes an aligned pillar-to-point projection module to improve the final prediction. Our anchor-free approach avoids hyperparameter search associated with past methods, simplifying 3D object detection while significantly improving upon state-of-the-art.