搜索医学 AI 论文与资源

ICLR 2026 Poster accepted paper at ICLR 2026. Correcting motion artifacts in scientific and medical imaging is important, as they significantly impact image quality. However, evaluating deep learning-based and classical motion correction methods remains fundamentally difficult due to the lack of accessible ground-truth target data. To address this challenge, we study three evaluation approaches: real-world evaluation based on reference scans, simulated motion, and reference-free evaluation, each with its merits and shortcomings. To enable evaluation with real-world motion artifacts, we release PMoC3D, a dataset consisting of unprocessed $\textbf{P}$aired $\textbf{Mo}$tion-$\textbf{C}$orrupted $\textbf{3D}$ brain MRI data.

ICLR 2026 Poster accepted paper at ICLR 2026. Arabic remains one of the most underrepresented languages in natural language processing research, particularly in medical applications, due to the limited availability of open-source data and benchmarks. The lack of resources hinders efforts to evaluate and advance the multilingual capabilities of Large Language Models (LLMs). In this paper, we introduce MedAraBench, a large-scale dataset consisting of Arabic multiple-choice question-answer pairs across various medical specialties. We constructed the dataset by manually digitizing a large repository of academic materials created by medical professionals in the Arabic-speaking region.

ICLR 2026 Poster accepted paper at ICLR 2026. The application of large language models (LLMs) in the medical field has garnered significant attention, yet their reasoning capabilities in more specialized domains like anesthesiology remain underexplored. To bridge this gap, we introduce AnesSuite, the first comprehensive dataset suite specifically designed for anesthesiology reasoning in LLMs. The suite features AnesBench, an evaluation benchmark tailored to assess anesthesiology-related reasoning across three levels: factual retrieval (System 1), hybrid reasoning (System 1.x), and complex decision-making (System 2). Alongside this benchmark, the suite includes three training datasets that provide an infrastructure for continued pre-training (CPT), supervised fine-tuning (SFT), and reinforcement learning with verifiable rewards (RLVR). Code/project link: https://github.com/MiliLab/AnesSuite

ICLR 2026 Poster accepted paper at ICLR 2026. Visual Question Answering (VQA) enables targeted and context-dependent analysis of medical images, such as chest X-rays (CXRs). However, existing VQA datasets for CXRs are typically constrained by simplistic and brief answer formats, lacking localization annotations (e.g., bounding boxes) and structured tags (e.g., region or radiological finding/disease tags). To address these limitations, we introduce MIMIC-Ext-CXR-QBA (abbr. CXR-QBA), a large-scale CXR VQA dataset derived from MIMIC-CXR, comprising 42 million QA-pairs with multi-granular, multi-part answers, detailed bounding boxes, and structured tags. Code/project link: https://github.com/philip-mueller/mimic-ext-cxr-qba/

ICLR 2026 Poster accepted paper at ICLR 2026. Current medical language model (LM) benchmarks often over-simplify the complexities of day-to-day clinical practice tasks and instead rely on evaluating LMs on multiple-choice board exam questions. In psychiatry especially, these challenges are worsened by fairness and bias issues, since models can be swayed by patient demographics even when those factors should not influence clinical decisions. Thus, we present an expert-created and annotated dataset spanning five critical domains of decision-making in mental healthcare: treatment, diagnosis, documentation, monitoring, and triage. This U.S. centric dataset — created without any LM assistance — is designed to capture the nuanced clinical reasoning and daily ambiguities mental health practitioners encounter, reflecting the inherent complexities of care delivery that are missing from existing datasets.

ICLR 2026 Poster accepted paper at ICLR 2026. Psychiatric comorbidity is clinically significant yet challenging due to the complexity of multiple co-occurring disorders. To address this, we develop a novel approach integrating synthetic patient electronic medical record (EMR) construction and multi-agent diagnostic dialogue generation. We create 502 synthetic EMRs for common comorbid conditions using a pipeline that ensures clinical relevance and diversity. Our multi-agent framework transfers the clinical interview protocol into a hierarchical state machine and context tree, supporting over 130 diagnostic states while maintaining clinical standards.

FORMED 将通用时间序列基础模型重用于医学时间序列分类，并通过任务相关通道嵌入、标签查询和共享解码注意力层，在不同医学时间序列数据集上进行轻量适配。

系统研究医学 MLLM 在医学图像视觉定位中的失效模式，提出 VGMED 评估数据集与 VGRefine 推理时方法，面向医学视觉问答与医学图像解释场景。

ICLR 2026 Poster 论文提出 MENTAT：一个由临床专家创建和标注、面向心理健康真实任务与模糊性的公平性评测数据集，用于评估语言模型在临床决策任务中的表现与偏差。

The PhysioNet/CinC Challenge 2012 dataset contains ICU time-series records used for mortality prediction and patient-specific outcome modeling. It remains a useful benchmark for clinical time-series modeling, missingness-aware learning, and early warning model development.

cMedQA2 is an updated Chinese community medical question answering dataset for question-answer matching and medical QA research. It is useful for training and evaluating Chinese medical retrieval, ranking, and answer selection models.

AMOS is an abdominal multi-organ segmentation benchmark with CT and MRI cases for evaluating versatile medical image segmentation models. It supports abdominal organ segmentation, modality-general segmentation, and benchmarking of robust 3D segmentation methods.

REFUGE is a retinal fundus imaging challenge dataset for glaucoma assessment. It supports glaucoma classification, optic disc and cup segmentation, fovea localization, and fair comparison of ophthalmology AI methods on color fundus photographs.

The RSNA Pneumonia Detection Challenge dataset is a chest radiograph benchmark for detecting pneumonia-related lung opacities. It supports object detection, chest X-ray classification, localization, and radiology AI evaluation under a competition framework.

MURA is a musculoskeletal radiograph dataset from Stanford for abnormality detection in upper extremity X-rays. It is used for radiology classification, fracture-related screening, musculoskeletal imaging AI, and human-AI comparison studies.

ACDC is a cardiac MRI dataset for automated cardiac diagnosis and segmentation. It supports left and right ventricular segmentation, myocardium segmentation, cardiac function quantification, and evaluation of robust cardiac image analysis methods.

UK Biobank Imaging provides large-scale imaging phenotypes linked to genetic, lifestyle, and health outcome data. It is used for population-scale medical imaging AI, disease risk prediction, representation learning, multimodal biomedical modeling, and epidemiological AI studies.

The Cancer Genome Atlas is a large cancer genomics resource with molecular, clinical, and pathology-related data across many cancer types. It is a foundation dataset for oncology AI, survival prediction, subtype discovery, multimodal cancer modeling, and translational biomarker research.

OASIS provides open-access neuroimaging datasets for studying normal aging, dementia, and brain structure. It is useful for brain MRI segmentation, age prediction, dementia classification, longitudinal modeling, and neuroimaging method benchmarking.

ADNI provides longitudinal neuroimaging, biomarker, clinical, and cognitive data for Alzheimer disease research. It supports disease progression modeling, dementia diagnosis, multimodal prediction, biomarker discovery, and clinical translation studies.

EchoNet-Dynamic is a cardiac ultrasound video dataset with expert annotations for left ventricular function. It is used for echocardiography video understanding, ejection fraction estimation, cardiac segmentation, and clinical video AI research.

CAMELYON17 is a digital pathology dataset for detecting breast cancer metastases in lymph node whole-slide images across multiple centers. It supports pathology classification, metastasis detection, weakly supervised learning, and domain generalization in histopathology AI.

The ISIC Archive is a large public dermatology image repository for skin lesion analysis. It is widely used for melanoma classification, lesion segmentation, dermoscopic image retrieval, bias and domain shift analysis, and clinical imaging benchmark development.

fastMRI is a raw MRI dataset for accelerated magnetic resonance image reconstruction, originally released by NYU Langone Health and Meta AI. It is used for MRI reconstruction, compressed sensing replacement, generative reconstruction, and robustness evaluation.

MedMNIST v2 is a standardized collection of lightweight biomedical image classification datasets, including 2D and 3D tasks. It is useful for quick benchmarking, AutoML, foundation model sanity checks, and reproducible evaluation across multiple medical imaging domains.

BraTS 2024 provides multimodal brain MRI data and expert annotations for brain tumor segmentation and related tumor subregion analysis. It is a major benchmark for glioma segmentation, radiology AI, and robust multimodal MRI segmentation methods.

C4KC-KiTS is a TCIA imaging collection associated with kidney and kidney tumor segmentation benchmarks. It supports kidney segmentation, renal tumor segmentation, surgical planning research, and evaluation of abdominal CT segmentation models.

LIDC-IDRI is a lung CT dataset with thoracic CT scans and expert nodule annotations. It is a classic benchmark for lung nodule detection, segmentation, malignancy characterization, radiomics, and computer-aided diagnosis research.

VinDr-CXR is a chest X-ray dataset with radiologist annotations from Vietnamese hospitals. It supports abnormality classification, lesion localization, radiology object detection, and robustness studies across clinical sites and populations.

NIH ChestX-ray14 is a public chest radiograph dataset with image-level labels for thoracic disease findings mined from reports. It is commonly used for chest X-ray classification, weak supervision, thoracic disease detection, and radiology benchmark comparisons.

CheXpert is a large chest radiograph dataset from Stanford with uncertainty-aware labels for common chest X-ray findings. It is widely used for radiology classification, label uncertainty modeling, chest X-ray representation learning, and clinical imaging benchmarks.

Sleep-EDF Expanded contains polysomnographic sleep recordings with EEG and related physiological signals. It is used for sleep stage classification, biosignal time-series modeling, self-supervised learning on physiological signals, and clinical sleep research benchmarks.

PTB-XL is a large public 12-lead electrocardiography dataset with diagnostic statements and waveform records. It is a standard benchmark for ECG classification, cardiac abnormality detection, clinical signal representation learning, and robust evaluation of biosignal models.

The eICU Collaborative Research Database is a multicenter critical care database containing deidentified ICU data from many hospitals. It is commonly used for external validation, ICU outcome prediction, temporal modeling, and cross-site generalization studies in clinical AI.

MIMIC-IV-ECG is a large deidentified electrocardiogram dataset linked to the MIMIC-IV clinical data ecosystem. It supports ECG classification, arrhythmia detection, representation learning, and multimodal modeling with structured EHR context.

MIMIC-CXR is a large deidentified chest radiograph dataset with associated free-text radiology reports. It is widely used for chest X-ray classification, report generation, image-text representation learning, radiology retrieval, and medical multimodal foundation model evaluation.

MIMIC-IV-Note provides deidentified clinical notes linked to MIMIC-IV hospital data. It supports clinical NLP tasks such as note representation learning, discharge summary modeling, information extraction, summarization, and multimodal EHR-text modeling.

MIMIC-IV is a large deidentified electronic health record dataset from Beth Israel Deaconess Medical Center, covering hospital and ICU data for critical care research. It is a core benchmark source for clinical prediction, temporal EHR modeling, phenotyping, and healthcare AI method development.

SLAKE is a semantically labeled medical visual question answering dataset with bilingual English-Chinese questions, medical images, and knowledge-enhanced annotations. It is useful for medical multimodal learning, image-grounded QA, and radiology VQA evaluation.

CMCQA is a large Chinese medical conversational question-answering dataset released with knowledge-grounded medical dialogue research. It supports medical conversation QA, knowledge-grounded response generation, and evaluation of Chinese medical dialogue systems.

HuatuoGPT2-SFT-GPT4-140K is a Chinese medical supervised fine-tuning dataset containing medical instruction-style conversations and GPT-4-assisted responses. It is useful for Chinese medical assistant alignment and medical LLM instruction tuning.

Huatuo-26M is a large-scale Chinese medical question-answering dataset with about 26 million QA pairs collected for medical language modeling and medical dialogue research. It is suitable for Chinese medical LLM pretraining, fine-tuning, and QA system development.

MedQA is a medical examination question answering benchmark with English and Chinese medical licensing-style question sets, including mainland China and Taiwan variants. It is widely used for medical QA and medical reasoning evaluation.

MedDG is an entity-centric Chinese medical consultation dataset with domain entity annotations for medical dialogue generation. It supports entity-aware response generation, medical consultation modeling, and dialogue systems that ground responses in clinical concepts.

CMB is a comprehensive Chinese medical benchmark for evaluating medical large language models on medical exams, reasoning, and clinical knowledge questions. It is suited for Chinese medical QA, LLM evaluation, and instruction-following assessment.

CBLUE is a Chinese biomedical language understanding benchmark covering real-world biomedical NLP tasks such as named entity recognition, relation extraction, term normalization, clinical trial classification, sentence similarity, and medical question answering. It is useful for evaluating Chinese clinical NLP models and medical language models.

Low-dose CT imaging collection from the National Lung Screening Trial, distributed by The Cancer Imaging Archive.

NIH Clinical Center chest X-ray dataset released for computer-aided detection and radiology machine learning research.