datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
multimodal-ct-radiology-reports
Perle AI Multi-phase CECT and CT with Radiology Reports
Summary
A de-identified CT dataset from Perle AI, paired with the original radiology reports. It supports work on multi-modal medical imaging: phase or pathology classification, report generation from images, and visual question answering.
The release has three configurations:
Config
Modality
Subjects
Pairing
cect_3phase
3-phase contrast-enhanced abdominal CT (DICOM)
5
per-subject text report +… See the full description on the dataset page: https://huggingface.co/datasets/Perle-ai/multimodal-ct-radiology-reports.ROCOv2-radiology
ROCOv2: Radiology Object in COntext version 2
Introduction
ROCOv2 is a multimodal dataset consisting of radiological images and associated medical concepts and captions extracted from the PMC Open Access Subset. It is an updated version of the ROCO dataset, adding 35,705 new images and improving concept extraction and filtering.
Dataset Overview
The ROCOv2 dataset contains 79,789 radiological images, each with a corresponding caption and medical concepts. The… See the full description on the dataset page: https://huggingface.co/datasets/eltorio/ROCOv2-radiology.Radiology_mini0.33% sampled from https://huggingface.co/datasets/eltorio/ROCOv2-radiology
ROCO-radiology
Dataset Summary
The "ROCO-radiology" dataset is derived from the Radiology Objects in COntext (ROCO) dataset, a large-scale medical and multimodal imaging collection. The language used is primarily English, and it covers the domain of medical imaging, specifically radiology. We only modified the dataset by choosing only for radiology dataset and convert the image into PIL Object. For further details and citation, pleaser refer to original author.
ROCO-radiologyThe "ROCO-radiology" dataset is derived from the Radiology Objects in COntext (ROCO) dataset, a large-scale medical and multimodal imaging collection. The language used is primarily English, and it covers the domain of medical imaging, specifically radiology. We only modified the dataset by choosing only for radiology dataset and convert the image into PIL Object. For further details and citation, pleaser refer to original author.… See the full description on the dataset page: https://huggingface.co/datasets/eltorio/ROCO-radiology.ROCO-radiology
Dataset Summary
The "ROCO-radiology" dataset is derived from the Radiology Objects in COntext (ROCO) dataset, a large-scale medical and multimodal imaging collection. The language used is primarily English, and it covers the domain of medical imaging, specifically radiology. We only modified the dataset by choosing only for radiology dataset and convert the image into PIL Object. For further details and citation, pleaser refer to original author.
radiology-conflictRadiology_Project_Annotatedelectronic-radiology-phd-thesis-trRROCOv2-radiology-minirocov2-questions-radiologyradiologyradiology-hu-asrROCO-radiologyROCO-radiology-50percentparrot-radiology-asr-en
PARROT Radiology ASR Dataset (Synthetic Speech)
Dataset Description
This dataset contains synthetic English radiology speech paired with transcriptions. It is designed for training and evaluating radiology-focused Automatic Speech Recognition models, speech LLMs, and multimodal medical AI systems. All audio is generated from the PARROT v1.0 radiology report corpus, a multilingual collection of fictional reports authored by expert radiologists from 21 countries.… See the full description on the dataset page: https://huggingface.co/datasets/ysdede/parrot-radiology-asr-en.RadFAQs_Radiology_Imaging_Health_QA_Dataset
RadFAQs — Radiology & Imaging Health Q&A Dataset (Nepali)
Overview
This dataset (radfaqs_health_qa_nepali_all_nepali.jsonl) is a large, single-source collection of 1,235 instruction-following conversation pairs in Nepali, entirely focused on radiology and medical imaging — CT scans, MRI, X-rays, ultrasound, PET scans, bone density (DEXA) tests, echocardiograms, angiography, radiotherapy/brachytherapy, and related procedures. Each record is a single-turn human↔gpt… See the full description on the dataset page: https://huggingface.co/datasets/sabin1234/RadFAQs_Radiology_Imaging_Health_QA_Dataset.Xray-OPG-Dental-Radiology-Reports-Without-Findings-DatasetDataset Description:
This dataset is a large-scale collection of OPG (Orthopantomogram) X-ray reports without clinical findings, containing data from 970 patients and 970 medical images, designed to support the development and training of advanced healthcare AI, medical imaging, diagnostic AI, and clinical NLP systems.
The dataset captures authentic imaging characteristics such as scanner variability, acquisition protocols, and patient positioning, along with clinical narratives. This ensures… See the full description on the dataset page: https://huggingface.co/datasets/InfoBayAI/Xray-OPG-Dental-Radiology-Reports-Without-Findings-Dataset.RadiologyX-Ray-Radiology-Reports-Without-Findings-DatasetDataset Description:
This dataset is a large-scale collection of X-ray (Radiography) reports without clinical findings, containing data from 199,352 patients and 380,512 medical images, designed to support the development and training of advanced healthcare AI, medical imaging, and diagnostic AI systems.
The dataset captures authentic imaging characteristics such as scanner variability, acquisition protocols, and patient positioning, along with clinical narratives. This ensures high-quality… See the full description on the dataset page: https://huggingface.co/datasets/InfoBayAI/X-Ray-Radiology-Reports-Without-Findings-Dataset.radiology-samples
Dataset Card for "radiology-samples"
More Information needed
ROCOv2-X-Ray-radiology_-cycle-1
ROCOv2 X-Ray, Report Generation Pilot (Cycle 1)
A 28-row pilot testing whether short ROCO captions can be expanded into report-shaped training pairs.
Why this exists
ROCOv2 captions are one or two clipped sentences, often written to make a teaching point rather than to read as a radiological finding. A vision-language model trained directly on them learns to produce clipped captions, not reports.
This pilot tested a different approach: take the caption and its… See the full description on the dataset page: https://huggingface.co/datasets/deepLEARNING786/ROCOv2-X-Ray-radiology_-cycle-1.ROCOv2-X-Ray-radiology
ROCOv2 X-Ray Subset
Radiographs extracted from ROCOv2 (Radiology Objects in COntext, version 2), for training vision-language models on X-ray interpretation.
What this is
ROCOv2 spans many imaging modalities. This subset keeps only the X-ray studies, so that a model can be trained on a single modality rather than learning across CT, MRI, ultrasound and radiography at once.
Rows
4,254
Split
train
Size
~977 MB
Modality
X-ray only… See the full description on the dataset page: https://huggingface.co/datasets/deepLEARNING786/ROCOv2-X-Ray-radiology.Chameleon-Radiology-Reportsradiology
Dataset Card for "radiology"
More Information needed
roco_v2_radiology_latents_sdxl_blip2CT-Scan-Radiology-Reports-Without-Findings-DatasetDataset Description:
This dataset is a large-scale collection of CT (Computed Tomography) scan reports without clinical findings, containing data from 41,669 patients and 12,715,504 medical images, designed to support the development and training of advanced healthcare AI, medical imaging, and diagnostic AI systems.
The dataset captures authentic imaging characteristics such as scanner variability, acquisition protocols, and patient positioning, along with structured and unstructured clinical… See the full description on the dataset page: https://huggingface.co/datasets/InfoBayAI/CT-Scan-Radiology-Reports-Without-Findings-Dataset.dental-radiologyradiology-coder-sft
Radiology Report ICD-10 Coder
Part of the AxisMapper Medical AI Suite — 16 domain-specific SFT datasets for fine-tuning medical LLMs.
Built by AmareshHebbar | Studio Ilios / Humanova Minds
What this dataset does
Radiology reports / impressions → ICD-10-CM codes for all documented findings
Why download this
Automate radiology coding, build report-to-code pipelines for RIS/PACS integration, or train models to extract diagnosis codes from chest X-ray… See the full description on the dataset page: https://huggingface.co/datasets/AmareshHebbar/radiology-coder-sft.radiology_audio_3_iphone_laptop_666_samples