Scale-Aware Domain Adaptation for Robust UAV Tracking

Review Geochemistry & Geophysics

Correlation Filters for Unmanned Aerial Vehicle-Based Aerial Tracking: A Review and Experimental Evaluation

Changhong Fu et al.

IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE (2022)

Article Computer Science, Artificial Intelligence

A Discriminative Single-Shot Segmentation Network for Visual Object Tracking

Alan Lukezic et al.

Summary: In this paper, a discriminative single-shot segmentation tracker called D3S(2) is proposed. It narrows the gap between visual object tracking and video object segmentation by applying two target models with complementary geometric properties. The overall tracking reliability is further increased by decoupling the object and feature scale estimation.

IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE (2022)

添加到收藏夹

Proceedings Paper Computer Science, Artificial Intelligence

Safe Self-Refinement for Transformer-based Domain Adaptation

Tao Sun et al.

Summary: In this paper, a novel method named SSRT (Safe Self-Refinement for Transformer-based domain adaptation) is proposed, which combines a transformer backbone with simple adversarial adaptation to achieve better performance than CNN on tasks with large domain gaps. A Safe Self-Refinement strategy is also proposed to reduce the risk of model collapse and improve knowledge transfer between domains.

2022 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR) (2022)

添加到收藏夹

Proceedings Paper Computer Science, Artificial Intelligence

Unsupervised Domain Adaptation for Nighttime Aerial Tracking

Junjie Ye et al.

Summary: This work presents an unsupervised domain adaptation framework for nighttime aerial tracking, which includes a unique object discovery approach and feature alignment layers. It also introduces a benchmark dataset for nighttime tracking. The experiments demonstrate the robustness and adaptability of the framework in nighttime aerial tracking.

2022 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR) (2022)

添加到收藏夹

Proceedings Paper Automation & Control Systems