FOCAL

FOCAL数据集是由佐治亚理工学院和福特汽车公司合作创建的，专注于视频主动学习领域的研究。该数据集包含126个视频序列，覆盖69个独特的城市场景，涉及多种天气、光照和季节条件。数据集旨在解决视频标注成本问题，通过提供详细的标注时间和成本信息，支持研究如何在预算限制下最大化性能。FOCAL数据集的应用领域包括自动驾驶和基础设施辅助自主系统，旨在通过分析视频序列中的交互和复杂性，优化标注策略和降低成本。

The FOCAL dataset is collaboratively created by the Georgia Institute of Technology and Ford Motor Company, focusing on research in the field of active learning for videos. This dataset contains 126 video sequences covering 69 unique urban scenarios, involving diverse weather, lighting and seasonal conditions. It aims to solve the problem of high video annotation costs, providing detailed annotation time and cost information to support research on maximizing performance under budget constraints. The application fields of the FOCAL dataset include autonomous driving and infrastructure-assisted autonomous systems, with the goal of optimizing annotation strategies and reducing costs by analyzing the interactions and complexities within video sequences.