STAIR动作描述视频字幕数据集

STAIR动作描述视频字幕数据集是由千叶工业大学STAIR实验室创建的，专注于描述人类动作的日语视频字幕数据集。该数据集包含79,822个视频和399,233条字幕，每个字幕详细描述了视频中的动作、人物和地点。数据集的创建过程涉及从现有视频数据集中收集视频，并通过众包服务进行多条字幕的标注。此数据集主要用于自动视频字幕生成，特别是在解决非英语语言资源稀缺问题方面具有重要意义，旨在通过提供详细的动作描述，增强视频内容的可搜索性和可理解性。

The STAIR Action Description Video Subtitle Dataset, created by the STAIR Laboratory of Chiba Institute of Technology, is a Japanese-language video subtitle dataset focused on describing human actions. This dataset contains 79,822 videos and 399,233 subtitle entries, with each subtitle elaborately describing the actions, characters, and locations featured in the corresponding video. The construction of this dataset involves collecting videos from existing video datasets and annotating multiple subtitle entries via crowdsourcing services. This dataset is primarily utilized for automatic video subtitle generation, and it holds significant importance in addressing the scarcity of non-English language resources. Its core goal is to enhance the searchability and comprehensibility of video content by providing detailed descriptions of human actions.