Related Experiment Video
Updated: Dec 2, 2025
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
Published on: November 30, 2022
Segmentation of coronary arteries images using global feature embedded network with active contour loss
Jia Gu1, Zhijun Fang1, Yongbin Gao1
1School of Electronic and Electrical Engineering, Shanghai University of Engineering Science, Shanghai, 201620, China.
Insights
This study introduces a new network to accurately segment coronary arteries in 3D CTA scans. The method improves noise reduction and boundary refinement for better coronary heart disease analysis.
Area of Science:
- Medical imaging analysis
- Cardiovascular disease research
- Artificial intelligence in healthcare
Background:
- Coronary heart disease (CHD) poses a significant global health threat with increasing morbidity and mortality.
- Accurate segmentation of coronary arteries in 3D coronary computed tomography angiography (CTA) data is complex and challenging due to image particularities.
Purpose of the Study:
- To develop an advanced deep learning framework for precise and efficient coronary artery segmentation in 3D CTA.
- To address the challenges of noise and boundary definition in medical image segmentation.
Main Methods:
- Proposed a novel global feature embedded network integrating multi-level network features for comprehensive semantic and detailed information.
- Incorporated improved noisy activating functions to mitigate noise artifacts in CTA data.
- Enhanced a learning active contour model for refined segmentation with smooth boundaries based on network-generated score maps.
Main Results:
- The proposed framework demonstrated state-of-the-art performance in coronary artery segmentation.
- Achieved intuitive and quantitative improvements in segmentation accuracy and boundary precision.
- Effectively reduced the impact of noise on segmentation outcomes.
Conclusions:
- The novel global feature embedded network offers a robust solution for accurate coronary artery segmentation in 3D CTA.
- The integrated noise reduction and boundary refinement techniques enhance the reliability of cardiovascular image analysis.
- This advancement holds potential for improved diagnosis and treatment planning for coronary heart disease.
Abstract:
Coronary heart disease (CHD) is a serious disease that endangers human health and life. In recent years, the morbidity and mortality of CHD are increasing significantly. Because of the particularity and complexity of medical image, it is challenging to segment coronary artery accurately and efficiently. This paper proposes a novel global feature embedded network for better coronary arteries segmentation in 3D coronary computed tomography angiography (CTA) data. The global feature combines multi-level layers from various stages of the network, which contains semantic information and detailed features, aiming to accurately segment target with precise boundary. In addition, we integrate a group of improved noisy activating functions with parameters into our network to eliminate the impact of noise in CTA data. And we improve the learning active contour model, which obtains a refined segmentation result with smooth boundary based on the high-quality score map produced by the networks. The experimental results show that the proposed framework achieved the state-of-the-art performance intuitively and quantitively.
