OmniMedVQA
收藏资源简介:
OmniMedVQA是一个专为医学领域设计的大型全面评估基准,包含12种不同模态和超过20个独特的人体解剖区域。该数据集由上海人工智能实验室创建,旨在解决现有医学图像数据集的多样性和覆盖范围不足的问题。OmniMedVQA包含118,010张真实医疗场景中的图像,确保与医学领域的要求和评估大型视觉语言模型(LVLMs)的适用性相符。数据集的创建过程涉及收集多个医学分类数据集,并利用强大的上下文推理能力将这些数据转换为视觉问答(VQA)格式。OmniMedVQA的应用领域广泛,旨在全面评估LVLMs在医学挑战中的基本能力,推动医学领域LVLMs的发展和应用。
OmniMedVQA is a large-scale comprehensive evaluation benchmark specifically designed for the medical domain, covering 12 distinct modalities and over 20 unique human anatomical regions. Developed by the Shanghai AI Laboratory, this dataset aims to address the shortcomings of existing medical image datasets in terms of diversity and coverage. OmniMedVQA contains 118,010 images from real medical scenarios, ensuring alignment with the requirements of the medical field and the suitability for evaluating Large Vision-Language Models (LVLMs). The dataset construction process involves collecting multiple medical classification datasets, and leveraging robust contextual reasoning capabilities to convert these data into Visual Question Answering (VQA) format. OmniMedVQA has a wide range of application scenarios, aiming to comprehensively evaluate the basic capabilities of LVLMs when facing medical challenges, and promote the development and application of LVLMs in the medical domain.




