OS VSD

OS VSD数据集是由IBM研究团队创建的，用于训练和测试场景变化检测算法。该数据集通过Yandex.Toloka平台收集，包含6871个场景变化，总时长达到965分钟，主要用于视频处理中的场景分割任务。数据集的创建过程包括使用多种算法配置以最大化结果的完整性，并通过众包方式进行标记，确保每个视频段至少由5人标记，以达成共识。该数据集的应用领域主要集中在视频分析和自动化视频处理技术中，旨在提高场景变化检测的准确性和效率。

OS VSD Dataset was created by the IBM Research team for training and testing scene change detection algorithms. Collected via the Yandex.Toloka platform, it comprises 6,871 scene change instances with a total duration of 965 minutes, and is primarily intended for scene segmentation tasks in video processing. The dataset creation process adopted multiple algorithm configurations to maximize the completeness of the results, and used crowdsourcing for annotation, ensuring that each video segment was labeled by at least five annotators to reach a consensus. The application domains of this dataset mainly focus on video analysis and automated video processing technologies, with the goal of improving the accuracy and efficiency of scene change detection.