OSP2B: One-Stage Point-to-Box Network for 3D Siamese Tracking
This work improves 3D object tracking for autonomous driving applications by offering a more efficient and aligned method, though it is incremental as it builds on the existing Siamese tracking paradigm.
The paper tackles the problem of 3D single object tracking in point clouds by proposing a one-stage point-to-box network to address hyper-parameter tuning and task misalignment in existing two-stage methods, achieving leading performance on benchmarks like KITTI and Waymo SOT Dataset with real-time speed.
Two-stage point-to-box network acts as a critical role in the recent popular 3D Siamese tracking paradigm, which first generates proposals and then predicts corresponding proposal-wise scores. However, such a network suffers from tedious hyper-parameter tuning and task misalignment, limiting the tracking performance. Towards these concerns, we propose a simple yet effective one-stage point-to-box network for point cloud-based 3D single object tracking. It synchronizes 3D proposal generation and center-ness score prediction by a parallel predictor without tedious hyper-parameters. To guide a task-aligned score ranking of proposals, a center-aware focal loss is proposed to supervise the training of the center-ness branch, which enhances the network's discriminative ability to distinguish proposals of different quality. Besides, we design a binary target classifier to identify target-relevant points. By integrating the derived classification scores with the center-ness scores, the resulting network can effectively suppress interference proposals and further mitigate task misalignment. Finally, we present a novel one-stage Siamese tracker OSP2B equipped with the designed network. Extensive experiments on challenging benchmarks including KITTI and Waymo SOT Dataset show that our OSP2B achieves leading performance with a considerable real-time speed.Code will be available at https://github.com/haooozi/OSP2B.