Back to datasets
Dataset assetOpen Source CommunityMedical ImagingTumor Detection
AbdomenAtlas 3.0
AbdomenAtlas 3.0 is the first publicly available high‑quality abdominal CT scan dataset with paired radiology reports. The database contains over 9,000 CT scans and their radiology reports, together with voxel‑level annotations of liver, kidney and pancreas tumors.
Source
github
Created
Nov 25, 2024
Updated
Dec 2, 2024
Signals
470 views
Availability
Linked source ready
Overview
Dataset description and usage context
RadGPT Dataset Overview
Dataset Introduction
- Name: AbdomenAtlas 3.0
- Type: Abdominal CT Scan Dataset
- Features: First publicly released high‑quality abdominal CT scan dataset with paired radiology reports.
Data Content
- Number of CT Scans: Over 9,000
- Number of Reports: Paired radiology reports
- Annotations: Voxel‑level annotations of liver, kidney and pancreas tumors
Model
- Model Name: RadGPT
- Function: Segmentation‑based report generation model
- Performance: Significantly outperforms current state‑of‑the‑art abdominal CT report generation models
Release Plan
- Status: Upcoming release
Need downstream help?
Pair the dataset with AI analysis and content workflows.
Once the source passes your review, move straight into summarization, transformation, report drafting, or presentation generation with the JuheAI toolchain.