Guardado en:
Detalles Bibliográficos
Autores principales: Sun, Haowei, Zhou, Kai, Gao, Hao, Zhang, Shiteng, Hu, Jinwu, Wen, Xutao, Ye, Qixiang, Tan, Mingkui
Formato: Preprint
Publicado: 2026
Materias:
Acceso en línea:https://arxiv.org/abs/2604.21453
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866918464202997760
author Sun, Haowei
Zhou, Kai
Gao, Hao
Zhang, Shiteng
Hu, Jinwu
Wen, Xutao
Ye, Qixiang
Tan, Mingkui
author_facet Sun, Haowei
Zhou, Kai
Gao, Hao
Zhang, Shiteng
Hu, Jinwu
Wen, Xutao
Ye, Qixiang
Tan, Mingkui
contents Visual Active Tracking (VAT) aims to control cameras to follow a target in 3D space, which is critical for applications like drone navigation and security surveillance. However, it faces two key bottlenecks in real-world deployment: confusion from visually similar distractors caused by insufficient instance-level discrimination and severe failure under occlusions due to the absence of active planning. To address these, we propose OA-VAT, a unified pipeline with three complementary modules. First, a training-free Instance-Aware Offline Prototype Initialization aggregates multi-view augmented features via DINOv3 to construct discriminative instance prototypes, mitigating distractor confusion. Second, an Online Prototype Enhancement Tracker enhances prototypes online and integrates a confidence-aware Kalman filter for stable tracking under appearance and motion changes. Third, an Occlusion-Aware Trajectory Planner, trained on our new Planning-20k dataset, uses conditional diffusion to generate obstacle-avoiding paths for occlusion recovery. Experiments demonstrate OA-VAT achieves 0.93 average SR on UnrealCV (+2.2% vs. SOTA TrackVLA), 90.8% average CAR on real-world datasets (+12.1% vs. SOTA GC-VAT), and 81.6% TSR on a DJI Tello drone. Running at 35 FPS on an RTX 3090, it delivers robust, real-time performance for practical deployment.
format Preprint
id arxiv_https___arxiv_org_abs_2604_21453
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Instance-level Visual Active Tracking with Occlusion-Aware Planning
Sun, Haowei
Zhou, Kai
Gao, Hao
Zhang, Shiteng
Hu, Jinwu
Wen, Xutao
Ye, Qixiang
Tan, Mingkui
Computer Vision and Pattern Recognition
Visual Active Tracking (VAT) aims to control cameras to follow a target in 3D space, which is critical for applications like drone navigation and security surveillance. However, it faces two key bottlenecks in real-world deployment: confusion from visually similar distractors caused by insufficient instance-level discrimination and severe failure under occlusions due to the absence of active planning. To address these, we propose OA-VAT, a unified pipeline with three complementary modules. First, a training-free Instance-Aware Offline Prototype Initialization aggregates multi-view augmented features via DINOv3 to construct discriminative instance prototypes, mitigating distractor confusion. Second, an Online Prototype Enhancement Tracker enhances prototypes online and integrates a confidence-aware Kalman filter for stable tracking under appearance and motion changes. Third, an Occlusion-Aware Trajectory Planner, trained on our new Planning-20k dataset, uses conditional diffusion to generate obstacle-avoiding paths for occlusion recovery. Experiments demonstrate OA-VAT achieves 0.93 average SR on UnrealCV (+2.2% vs. SOTA TrackVLA), 90.8% average CAR on real-world datasets (+12.1% vs. SOTA GC-VAT), and 81.6% TSR on a DJI Tello drone. Running at 35 FPS on an RTX 3090, it delivers robust, real-time performance for practical deployment.
title Instance-level Visual Active Tracking with Occlusion-Aware Planning
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2604.21453