DA-Tracker

DA-Tracker数据集是由莫纳什大学数据科学与人工智能系创建，包含57个视频序列，总计超过30,000帧，涵盖两种不同的蚂蚁种类在不同背景下的移动轨迹。该数据集用于评估无监督领域适应框架在多目标跟踪中的应用，旨在解决不同蚂蚁种类和环境下的自动跟踪问题。数据集创建过程包括使用深度学习模型进行初步跟踪，随后由专家手动校正，确保高质量的跟踪结果。

The DA-Tracker Dataset was developed by the Department of Data Science and Artificial Intelligence at Monash University. It includes 57 video sequences with a total of over 30,000 frames, covering the movement trajectories of two distinct ant species under varying background conditions. This dataset is utilized to evaluate unsupervised domain adaptation frameworks for multi-object tracking, aiming to resolve the automatic tracking issue across different ant species and environmental contexts. The dataset creation process involved initial tracking using deep learning models, followed by manual correction conducted by domain experts to ensure high-quality tracking results.