Related Experiment Video
Updated: Mar 30, 2026
05:24
Multifractal Spectrum Analysis for Assessing Pulmonary Nodule Malignancy
Published on: January 10, 2025
1.0K
Unsupervised segmentation of lung fields in chest radiographs using multiresolution fractal feature vector and
Wen-Li Lee1, Koyin Chang2, Kai-Sheng Hsieh3
1Department of Healthcare Information and Management, Ming Chuan University, Taoyuan, 333, Taiwan, ROC. wllee2002@msn.com.
Medical & Biological Engineering & Computing
|November 5, 2015
Summary
This study introduces an unsupervised method for segmenting lung fields in chest radiographs using fractal features and fuzzy clustering. This technique accurately identifies lung regions, enabling reliable measurement of the cardiothoracic ratio (CTR) for cardiac health assessment.
Area of Science:
- Medical Imaging
- Computer Vision
- Radiology
Background:
- Accurate lung field segmentation in chest radiographs is crucial for automated image analysis.
- Existing methods may lack robustness or require supervision.
- The cardiothoracic ratio (CTR) is a key indicator for assessing cardiac hypertrophy.
Purpose of the Study:
- To develop an unsupervised method for precise lung field segmentation in chest radiographs.
- To evaluate the effectiveness of a novel multiresolution fractal feature vector approach.
- To demonstrate the utility of accurate lung segmentation in calculating the cardiothoracic ratio (CTR).
Main Methods:
- An unsupervised segmentation approach utilizing a multiresolution fractal feature vector.
- Application of fuzzy c-means clustering for initial contour generation.
- Refinement of contours using deformable models for precise lung field delineation.
Main Results:
- The proposed method effectively characterizes lung field regions.
- Satisfactory initial contours were obtained using fuzzy c-means clustering.
- The technique demonstrated high performance and feasibility for lung field segmentation.
Conclusions:
- The developed unsupervised method offers a robust solution for lung field segmentation in chest radiographs.
- Accurate lung segmentation facilitates reliable cardiothoracic ratio (CTR) measurement.
- This approach can aid physicians in identifying potential cardiac issues and guiding further diagnostic tests.
