Level Set Image Feature Detection and Application in COVID-19 Image Feature Knowledge Detection

Area of Science:

• Medical Imaging Analysis
• Artificial Intelligence in Healthcare
• Image Segmentation Techniques

Background:

• AI models show promise in medical image analysis, including COVID-19 detection.
• Robustness issues persist in AI segmentation for images with non-uniform density or multiphase targets.
• The Chan-Vese (CV) model is a representative image segmentation method.

Purpose of the Study:

• To evaluate a recent level set (LV) model for medical image segmentation.
• To assess the performance of a filtering variational method in detecting target characteristics.
• To address the limitations of existing AI models in handling complex medical imaging data.

Main Methods:

• Utilized a filtering variational method based on global medical pathology.
• Applied a recent level set (LV) model for image segmentation.
• Compared the proposed method's feature extraction capability against other LV models.

Main Results:

• The filtering variational method demonstrated superior image feature quality compared to other LV models.
• The proposed LV algorithm effectively detected lung region features in COVID-19 images.
• The algorithm showed adaptability across diverse medical imaging datasets.

Conclusions:

• The proposed LV method, using a filtering variational approach, offers robust performance in medical image analysis.
• This technique improves upon existing models for segmenting complex or non-uniform medical images.
• The LV method is a promising clinically adjunctive tool for machine-learning healthcare models.