UIT-ViIC

UIT-ViIC是越南信息科技大学创建的首个越南语图像描述数据集，包含3,850张与球类运动相关的图像，共有19,250条越南语描述。数据集内容主要来源于Microsoft COCO数据集，专注于球类运动图像，通过人工编写描述来确保语言的自然性和准确性。创建过程中，使用了专门的注释工具和严格的注释指南，确保描述的质量。该数据集主要应用于图像描述生成领域，旨在解决越南语环境下图像描述自动生成的难题。

UIT-ViIC is the first Vietnamese image captioning dataset developed by the University of Information Technology (UIT), Vietnam. It contains 3,850 images related to ball sports, with a total of 19,250 Vietnamese captions. The dataset is primarily sourced from the Microsoft COCO dataset, focusing on ball sports images, and its captions are manually written to guarantee linguistic naturalness and accuracy. During the dataset construction, specialized annotation tools and strict annotation guidelines were adopted to ensure the quality of the captions. This dataset is mainly applied in the field of image caption generation, aiming to address the challenges of automatic image caption generation in Vietnamese contexts.