GMAI-MMBench
收藏资源简介:
GMAI-MMBench是由上海人工智能实验室等机构创建的综合性医疗AI评估基准,包含285个高质量数据集,覆盖39种医疗图像模态和18个临床任务。数据集内容丰富,包括2D检测、分类和2D/3D分割等多种任务,数据来源于全球各地的公共和医院资源。创建过程中,数据集经过严格筛选和标准化处理,确保了数据的多样性和临床相关性。该数据集主要用于评估和提升大型视觉语言模型在医疗领域的应用,特别是在疾病诊断和治疗方面的辅助能力。
GMAI-MMBench is a comprehensive medical AI evaluation benchmark developed by institutions including Shanghai AI Laboratory. It comprises 285 high-quality datasets, covering 39 medical imaging modalities and 18 clinical tasks. The benchmark encompasses diverse tasks such as 2D detection, classification, and 2D/3D segmentation, with data sourced from global public and hospital medical resources. During its curation, the dataset has undergone rigorous screening and standardization processes to ensure its diversity and clinical relevance. This benchmark is primarily used to evaluate and enhance the applications of large vision-language models in the medical field, particularly their auxiliary capabilities in disease diagnosis and treatment.

- 1GMAI-MMBench: A Comprehensive Multimodal Evaluation Benchmark Towards General Medical AI上海人工智能实验室 · 2024年



