数据集导航

医疗数据集导航 - 全部数据集（第 162 页）

汇集 20 大疾病系统 5035 个医疗数据集，支持搜索与分类筛选，助力医疗AI研发选型

共 5035 个数据集 · 第 162 / 168 页
数据集名称 系统分类 病种 数据模态 任务类型 描述 来源 原始地址 操作
AnesSuite医学文本与大模型文本麻醉学问答4427 anesthesiology MCQ items focused on complex decision-making, specialized knowledge QA 2025GitHubhttps://arxiv.org/pdf/2504.02404访问
TracSum医学文本与大模型文本医疗摘要500 abstracts resulting in 3500 summary-citation traceable pairs, aspect-based summarization EMNLP 2025GitHubhttps://arxiv.org/pdf/2508.13798访问
MedAgentBoard通用医学数据集多模态多智能体协作8 benchmark categories for multi-agent reasoning tasks, text/image/EHR multi-agent collaboration 2025Project Pagehttps://arxiv.org/pdf/2505.12371访问
HEAL-MedVQA医学文本与大模型多模态视觉问答11000+ samples requiring localization prior to medical answering, grounded medical VQA IJCAI 2025Project Pagehttps://arxiv.org/pdf/2505.00744访问
BRIDGE医学文本与大模型文本多任务评估1.4 million samples covering 87 tasks in 9 languages, real-world clinical practice text 2025Project Pagehttps://arxiv.org/pdf/2504.19467访问
LLMEval-Med医学文本与大模型文本临床问答验证~1000 real-world cases validated via physician-in-the-loop audits, clinical QA validation 2025GitHubhttps://arxiv.org/pdf/2506.04078访问
CSEDB医学文本与大模型文本安全性评估30 criteria across 26 specialties based on expert physician consensus, safety-effectiveness eval 2025GitHubhttps://arxiv.org/pdf/2507.23486访问
SSG-VQA医学文本与大模型多模态手术视觉问答1300+ scene graph samples focused on instrument-tissue interaction, surgical VQA 2025GitHubhttps://arxiv.org/pdf/2506.06232访问
PET2Rep医学文本与大模型多模态报告生成565 whole-body paired PET/CT data combinations with detailed radiology reports, report generation 2025GitHubhttps://arxiv.org/pdf/2508.04062访问
MedTVT-QA医学文本与大模型多模态视觉问答3232 VQA pairs encompassing 15 medical specialties, 3 categories and 8 sub-categories clinical tasks ACL 2025GitHubhttps://arxiv.org/pdf/2506.18512访问
MediConfusion医学文本与大模型多模态视觉问答176 confusing pairs of two images sharing same question but different correct answers, VQA reliability ICLR 2025Hugging Facehttps://arxiv.org/abs/2409.15477访问
GMAI-MMBench医学文本与大模型多模态视觉问答26K QA pairs, 38 modality types, comprehensive multimodal evaluation benchmark NeurIPS 2024Hugging Facehttps://huggingface.co/datasets/OpenGVLab/GMAI-MMBench访问
PathMMU医学文本与大模型多模态病理推理33428 QAs, 24067 images, massive multimodal expert-level pathology understanding and reasoning 2024Hugging Facehttps://huggingface.co/datasets/jamessyx/PathMMU访问
CARES眼病多模态可信度评估41K QA pairs, trustworthiness benchmark for medical vision language models, open/closed QA NeurIPS 2024GitHubhttps://github.com/richard-peng-xia/CARES访问
MultiMedEval眼病多模态多任务评估6 tasks, 23 datasets, benchmark and toolkit for evaluating medical vision-language models 2024GitHubhttps://github.com/corentin-ryr/MultiMedEval访问
medical-o1-reasoning-SFT医学文本与大模型文本医学推理19.7k QA pairs, HuatuoGPT-o1 medical complex reasoning, medical VQA & reasoning ACL 2025Hugging Facehttps://huggingface.co/datasets/FreedomIntelligence/medical-o1-reasoning-SFT访问
ReasonMed医学文本与大模型文本医学推理370K multi-agent generated dataset from 1.75M CoT paths, medical reasoning, QA, chain-of-thought 2025Hugging Facehttps://arxiv.org/abs/2506.09513访问
Lingshu Train通用医学数据集多模态多模态训练~9.3M training samples from 60+ datasets, generalist foundation model for unified multimodal medical understanding 2025Project Pagehttps://arxiv.org/pdf/2506.07044访问
MedEvalKit Lingshu test通用医学数据集多模态基准评测152066 evaluation samples from 16 benchmark datasets, VQA, report generation, medical text QA 2025GitHubhttps://arxiv.org/pdf/2507.04289访问
MIRIAD医学文本与大模型文本医学问答5.82M/4.48M medical query-response pairs, augmenting LLMs with medical query-response, RAG & hallucination detection 202Hugging Facehttps://arxiv.org/abs/2506.06091访问
ClinBench-HPB医学文本与大模型文本临床问答3535 MCQs and 337 clinical cases covering 465+ hepato-pancreato-biliary diseases 2025Project Pagehttps://arxiv.org/abs/2506.00095访问
SurgVLM-DB眼病视频手术问答1.81M frames, 7.79M dialogues, large-scale multimodal surgical database 2025GitHubhttps://github.com/jinlab-imvr/SurgVLM访问
EndoBench消化系统疾病多模态内镜分析6832 clinically validated VQA samples, gastroscopy/colonoscopy/capsule/surgical endoscopy, 12 tasks 2025Hugging Facehttps://arxiv.org/abs/2505.23601v1访问
MedXpertQA医学文本与大模型多模态专家级问答4460 questions (text 2455 / image 2005), expert-level medical reasoning and understanding ICML 2025Hugging Facehttps://huggingface.co/datasets/TsinghuaC3I/MedXpertQA访问
MedCaseReasoning医学文本与大模型文本诊断推理14489 QA pairs, diagnostic reasoning from clinical case reports 2025GitHubhttps://github.com/kevinwu23/Stanford-MedCaseReasoning访问
DrVD-Bench眼病多模态诊断推理7789 image-text QA pairs, vision-language models reasoning like human doctors in medical image diagnosis 2025GitHubhttps://github.com/Jerry-Boss/DrVD-Bench访问
MedS-Ins医学文本与大模型文本指令微调5M samples, 19K instructions, versatile LLMs for medicine instruction tuning 2025Hugging Facehttps://huggingface.co/datasets/Henrychur/MedS-Ins访问
MedS-Bench医学文本与大模型文本临床任务评估11 categories of clinical tasks, benchmark for medical LLMs 2025Hugging Facehttps://huggingface.co/datasets/Henrychur/MedS-Bench访问
MM-Skin皮肤与结缔组织疾病多模态皮肤病问答Image-text dataset derived from textbooks for dermatology VLM enhancement 2025GitHubhttps://github.com/ZwQ803/MM-Skin访问
AlphaMed19K医学文本与大模型文本医学推理19K QA pairs, medical LLM reasoning with minimalist rule-based RL 2025Hugging Facehttps://huggingface.co/datasets/che111/AlphaMed19K访问