P2B: Point-to-Box Network for 3D Object Tracking in Point Clouds

TL;DR

P2B introduces an end-to-end point-to-box network for efficient 3D object tracking in point clouds, combining target localization and verification to outperform existing methods with real-time speed.

Contribution

The paper presents a novel point-to-box network that localizes and verifies 3D targets jointly, reducing search time and improving accuracy in point cloud tracking.

Findings

Achieves ~10% improvement over state-of-the-art on KITTI dataset.

Runs at 40 FPS on a single NVIDIA 1080Ti GPU.

Effective in real-time 3D object tracking.

Abstract

Towards 3D object tracking in point clouds, a novel point-to-box network termed P2B is proposed in an end-to-end learning manner. Our main idea is to first localize potential target centers in 3D search area embedded with target information. Then point-driven 3D target proposal and verification are executed jointly. In this way, the time-consuming 3D exhaustive search can be avoided. Specifically, we first sample seeds from the point clouds in template and search area respectively. Then, we execute permutation-invariant feature augmentation to embed target clues from template into search area seeds and represent them with target-specific features. Consequently, the augmented search area seeds regress the potential target centers via Hough voting. The centers are further strengthened with seed-wise targetness scores. Finally, each center clusters its neighbors to leverage the ensemble power for joint 3D target proposal and verification. We apply PointNet++ as our backbone and experiments on KITTI tracking dataset demonstrate P2B's superiority (~10%'s improvement over state-of-the-art). Note that P2B can run with 40FPS on a single NVIDIA 1080Ti GPU. Our code and model are available at https://github.com/HaozheQi/P2B.