MSVD-Turkish

MSVD-Turkish是一个大规模的视频描述数据集，由哈塞特佩大学计算机工程系创建。该数据集包含约2000个视频片段，总计80000条土耳其语视频描述。这些描述是通过仔细翻译MSVD数据集中的英语描述得到的。MSVD-Turkish不仅支持土耳其语视频字幕的研究，还因其英土双语描述，适用于多模态机器翻译研究。数据集的创建过程中，首先使用Google翻译API进行自动翻译，随后由双语专家进行校对和修正，确保描述的准确性和质量。该数据集的应用领域包括视频字幕生成和多模态机器翻译，旨在解决非英语语言的视频理解和描述问题。MSVD-Turkish is a large-scale video caption dataset created by the Department of Computer Engineering at Hacettepe University. This dataset contains approximately 2000 video clips, with a total of 80,000 Turkish video captions. These descriptions are developed via meticulous translation of the English captions from the original MSVD dataset. MSVD-Turkish not only supports research on Turkish video captioning, but also is applicable to multimodal machine translation research due to its English-Turkish bilingual captions. During the dataset construction process, automatic translation was first carried out using the Google Translate API, followed by proofreading and revision by bilingual experts to guarantee the accuracy and quality of the captions. The application fields of this dataset cover video caption generation and multimodal machine translation, aiming to address the challenges of video understanding and description for non-English languages.