Abstract: Current multimodal sensor trackers mainly rely on visual cues for target tracking. However, in challenging scenarios, the visual information acquired by multimodal sensors has limited ...

Abstract: Existing visual trackers based on Siamese region proposal networks determine the target location and size by comparing features of the target template and search regions. However, due to ...