TableVQA-Bench

TableVQA-Bench是由NAVER 云 AI创建的一个视觉问答基准数据集，专注于多个表格领域。该数据集包含1500个问答对，来源于现有的表格问答和表格结构识别数据集。数据集的创建过程包括使用样式表或自研的表格渲染系统获取图像，以及利用大型语言模型生成问答对。TableVQA-Bench的应用领域主要集中在视觉表格数据的问答任务，旨在解决现有数据集缺乏图像和问答对的问题，从而推动多模态大型语言模型在视觉表格理解任务中的性能评估和改进。

TableVQA-Bench is a visual question answering (VQA) benchmark dataset developed by NAVER Cloud AI, focusing on multiple table-related domains. This dataset includes 1500 question-answer pairs sourced from existing table question answering and table structure recognition datasets. The dataset creation workflow involves acquiring table images via style sheets or a self-developed table rendering system, as well as generating question-answer pairs using large language models (LLMs). The primary application scenarios of TableVQA-Bench are visual table data question answering tasks, and it aims to address the shortage of paired images and question-answer samples in existing datasets, thereby promoting the performance evaluation and improvement of multimodal large language models in visual table understanding tasks.