Coronary Calcium Detection Based on Improved Deep Residual Network in Mimics

Area of Science:

• Medical image processing
• Artificial intelligence in healthcare
• Cardiovascular imaging

Background:

• Coronary calcium detection is crucial for cardiovascular risk assessment.
• Low resolution and complex backgrounds in medical images pose challenges for accurate detection.
• Existing algorithms may struggle with accuracy and efficiency.

Purpose of the Study:

• To propose an improved coronary calcium detection algorithm for medical images.
• To enhance the accuracy and efficiency of coronary calcium detection.
• To address limitations of existing detection methods.

Main Methods:

• An improved Single Shot MultiBox Detector (SSD) algorithm was developed.
• Aggregate channel feature model was used for image preprocessing.
• Resnet-50 replaced VGG-16 to improve deep network performance.
• A two-parameter loss function was employed for network optimization.

Main Results:

• The proposed algorithm demonstrated superior performance compared to existing methods.
• Detection efficiency was significantly improved.
• Accuracy of coronary calcium detection was maintained at a high level.

Conclusions:

• The improved SSD algorithm offers enhanced accuracy and efficiency for coronary calcium detection in medical images.
• The algorithm shows promise for clinical applications in cardiovascular risk assessment.
• Deep learning approaches can effectively overcome challenges in medical image analysis.