F-Siamese Tracker: A Frustum-based Double Siamese Network for 3D Single Object Tracking

Abstract

This paper presents F-Siamese Tracker, a novel approach for single object tracking prominently characterized by more robustly integrating 2D and 3D information to reduce redundant search space. A main challenge in 3D single object tracking is how to reduce search space for generating appropriate 3D candidates. Instead of solely relying on 3D proposals, firstly, our method leverages the Siamese network applied on RGB images to produce 2D region proposals which are then extruded into 3D viewing frustums. Besides, we perform an online accuracy validation on the 3D frustum to generate refined point cloud searching space, which can be embedded directly into the existing 3D tracking backbone. For efficiency, our approach gains better performance with fewer candidates by reducing search space. In addition, benefited from introducing the online accuracy validation, for occasional cases with strong occlusions or very sparse points, our approach can still achieve high precision, even when the 2D Siamese tracker loses the target. This approach allows us to set a new state-of-the-art in 3D single object tracking by a significant margin on a sparse outdoor dataset (KITTI tracking). Moreover, experiments on 2D single object tracking show that our framework boosts 2D tracking performance as well.

Model-free Vehicle Tracking and State Estimation in Point Cloud Sequences

“…In the era of deep learning, the learned neural networks replace the manually designed models. P2B [27] generates and verifies bounding boxes guided by a learned PointNet++ [26], Zou et al [40] manages to fuse RGB information using frustums, Zarzar et al [38] refines the detection results by association, while [36], [19] apply convolutional neural networks(CNN) on the Bird-Eye-View images of LiDAR, and Giancola et al [11] learns to complete the partial point cloud then searches for it.…”