Related Experiment Video
Updated: Jun 23, 2026
Functional Neuroimaging Using Ultrasonic Blood-brain Barrier Disruption and Manganese-enhanced MRI
Published on: July 12, 2012
Duke Liver Dataset: A Publicly Available Liver MRI Dataset with Liver Segmentation Masks and Series Labels.
Jacob A Macdonald1, Zhe Zhu1, Brandon Konkel1
1From the Department of Radiology (J.A.M., Z.Z., B.K., M.A.M., W.F.W., M.R.B.), Department of Electrical and Computer Engineering (M.A.M.), Department of Computer Science (M.A.M.), Center for Advanced Magnetic Resonance Development (M.R.B.), and Division of Gastroenterology, Department of Medicine (M.R.B.), Duke University, 2301 Erwin Rd, Durham, NC 27710.
The Duke Liver Dataset offers 2146 abdominal MRI scans from 105 patients, featuring many with cirrhosis. It includes 310 segmented liver masks for research.
Area of Science:
- Radiology and Medical Imaging
- Hepatology
- Medical Image Analysis
Background:
- Liver disease, particularly cirrhosis, poses a significant global health challenge.
- Accurate imaging biomarkers are crucial for diagnosing and monitoring liver conditions.
- Abdominal Magnetic Resonance Imaging (MRI) is a key modality for liver assessment.
Purpose of the Study:
- To introduce and describe the Duke Liver Dataset, a valuable resource for liver imaging research.
- To provide a comprehensive collection of abdominal MRI series for the study of liver diseases.
- To facilitate the development and validation of automated liver segmentation techniques.
Main Methods:
- The dataset comprises 2146 abdominal MRI series.
- Data was collected from 105 patients, with a focus on those exhibiting cirrhotic features.
- A subset of 310 image series includes corresponding manually segmented liver masks.
Main Results:
- The dataset represents a substantial collection of abdominal MRI data.
- It includes a significant number of cases with liver cirrhosis.
- The availability of manual liver segmentations enables supervised learning approaches.
Conclusions:
- The Duke Liver Dataset is a valuable resource for advancing liver disease research using medical imaging.
- It supports the development of AI-driven tools for liver analysis and segmentation.
- This dataset can accelerate research in hepatology and radiological diagnostics.
