Related Experiment Video
Updated: Oct 29, 2025
10:39
A Label-Free Segmentation Approach for Intravital Imaging of Mammary Tumor Microenvironment
Published on: May 24, 2022
2.5K
X-Ray cardiac angiographic vessel segmentation based on pixel classification using machine learning and region
E O Rodrigues1, L O Rodrigues2, J J Lima1
1Department of Academic Informatics (DAINF), Universidade Tecnologica Federal do Parana (UTFPR), Pato Branco, Parana, Brazil.
Biomedical Physics & Engineering Express
|July 13, 2021
Summary
This study introduces a novel pixel-classification method for segmenting vessels in X-ray angiograms, achieving 95.48% accuracy. The approach utilizes textural features and a region-growing technique, outperforming existing unsupervised methods.
Area of Science:
- Medical Imaging
- Computer Vision
- Image Analysis
Background:
- Accurate vessel segmentation in X-ray angiograms is crucial for diagnosing various cardiovascular conditions.
- Existing unsupervised methods often struggle to achieve high accuracy and robustness.
Purpose of the Study:
- To develop and evaluate a novel pixel-classification approach for precise vessel segmentation in X-ray angiograms.
- To improve the accuracy and performance of automated vessel segmentation compared to state-of-the-art unsupervised methods.
Main Methods:
- Extraction of textural features (anisotropic diffusion, Hessian matrix, mathematical morphology, statistics) from pixel neighborhoods.
- Implementation of the ELEMENT methodology, a region-growing-controlled pixel classification that iteratively refines results.
- Utilization of the Random Forests classifier for pixel-level vessel structure prediction.
Main Results:
- The proposed pixel-classification approach achieved a state-of-the-art accuracy of 95.48% for vessel segmentation.
- The method demonstrated superior performance compared to existing unsupervised segmentation techniques.
- The ELEMENT methodology effectively enhanced classification accuracy through iterative refinement.
Conclusions:
- The developed pixel-classification approach offers a highly accurate and effective solution for vessel segmentation in X-ray angiograms.
- This method holds significant potential for improving the diagnosis and monitoring of vascular diseases.
- The combination of advanced textural features and the ELEMENT methodology represents a promising direction for medical image analysis.
