Related Experiment Video
Updated: Oct 22, 2025
08:41
Novel In Vivo Micro-Computed Tomography Imaging Techniques for Assessing the Progression of Non-Alcoholic Fatty Liver Disease
Published on: March 24, 2023
1.4K
Liver Segmentation in MRI Images using an Adaptive Water Flow Model.
Marjan Heidari1, Mehdi Taghizadeh2, Hassan Masoumi3
1PhD candidate, Department of Biomedical Engineering, Kazerun Branch, Islamic Azad University, Kazerun, Iran.
Journal of Biomedical Physics & Engineering
|August 30, 2021
Summary
This study introduces an automatic liver segmentation algorithm using an adaptive water flow model and a Multi-Layer Perception neural network for enhanced computer-aided diagnosis in MRI scans.
Area of Science:
- Medical Imaging
- Computer-Aided Diagnosis
- Image Segmentation
Background:
- Accurate liver segmentation is crucial for surgical planning and treating liver diseases.
- Current computer-aided diagnosis methods for liver segmentation face challenges.
- Precise localization of liver surface and segments is essential for effective treatment.
Purpose of the Study:
- To develop an automatic liver segmentation algorithm using an adaptive water flow model.
- To enhance liver region extraction through a probability distribution function-based transfer function.
- To perform final segmentation via a classification algorithm for improved accuracy.
Main Methods:
- An automatic liver segmentation algorithm was developed for Magnetic Resonance Imaging (MRI).
- A transfer function enhanced liver areas based on pixel probability distribution.
- An adaptive water flow model segmented the enhanced image, controlled by liver location and pixel gray levels.
- A Multi-Layer Perception (MLP) neural network classified candidate liver segments using texture, area, and gray level features.
Main Results:
- The algorithm achieved perfect liver segmentation on 250 MRI test images.
- It successfully distinguished the liver from surrounding organs.
- Quantitative evaluation demonstrated 97% accuracy, outperforming other evaluated algorithms.
Conclusions:
- Adaptive water flow and MLP classification provide robust liver segmentation in MRI.
- This method offers more reliable results compared to traditional pixel classification.
- The developed algorithm enhances computer-aided diagnosis for liver diseases.
