Related Experiment Video
Updated: Mar 31, 2026
08:21
Use of 3D Robotic Ultrasound for In Vivo Analysis of Mouse Kidneys
Published on: August 12, 2021
4.1K
Fully automatized renal parenchyma volumetry using a support vector machine based recognition system for
Oliver Gloger1, Klaus Tönnies, Birger Mensel
1Ernst Moritz Arndt University of Greifswald, Institute for Community Medicine, Walther-Rathenau-Str. 48. 17475 Greifswald, Germany.
Physics in Medicine and Biology
|October 29, 2015
Summary
This study introduces an automated framework for precise kidney segmentation in MRI data, improving organ volumetry for large population studies. The novel approach enhances accuracy and reduces variability compared to manual methods.
Area of Science:
- Medical Imaging
- Biomedical Engineering
- Radiology
Background:
- Medical image data, particularly MRI, has significantly increased, necessitating efficient organ segmentation for applications like organ volumetry.
- Manual segmentation is time-consuming and suffers from reader variability, posing challenges for large-scale studies.
- Automatic organ segmentation in native MR data is difficult due to imaging artifacts and intensity variations.
Purpose of the Study:
- To develop a robust, automated framework for accurate renal parenchyma segmentation in MR volume data.
- To improve organ volumetry in large-scale population studies by overcoming limitations of manual segmentation.
- To create subject-specific probability maps for renal tissue and exclude renal cysts.
Main Methods:
- A modularized, two-stepped probabilistic approach was developed.
- A three-class support vector machine (SVM) system incorporating Fourier descriptors was used for characteristic parenchyma part recognition.
- Probabilistic methods generated subject-specific parenchyma probability maps, refined by 3D level set segmentation and cyst exclusion.
Main Results:
- The framework successfully generated subject-specific probability maps for renal parenchyma.
- It accurately segmented renal parenchyma and excluded renal cysts, crucial for renal function analysis.
- Quantitative evaluation using volume errors and Dice coefficients demonstrated superior performance over existing methods.
Conclusions:
- The proposed framework offers a significant advancement in automated organ segmentation for MR imaging.
- It provides a reliable and accurate method for renal parenchyma volumetry in large studies.
- This automated approach addresses the challenges of manual segmentation, paving the way for more efficient medical research and clinical practice.
