Automatic Segmentation of Lumen Intima Layer in Longitudinal Mode Ultrasound Images.

Area of Science:

• Medical Imaging
• Biomedical Engineering
• Image Processing

Background:

• Accurate segmentation of the lumen intima layer in common carotid artery ultrasound images is crucial for cardiovascular disease assessment.
• Existing methods may lack the precision or automation required for widespread clinical application.

Purpose of the Study:

• To develop and validate an automated hybrid method for precise lumen intima layer segmentation in longitudinal mode ultrasound images of the common carotid artery.

Main Methods:

• A hybrid approach combining an active oblong optimization for coarse segmentation with post-processing (median filtering, Canny edge detection, curve fitting) for fine delineation.
• The method optimizes a locally defined contrast function of an active oblong with five degrees-of-freedom.
• Post-processing refines segmentation within the annulus region identified by the active oblong.

Main Results:

• The algorithm achieved an average accuracy of 98.9% on 84 ultrasound images.
• A Dice similarity index of 95.2% was obtained, indicating excellent overlap between segmented and true boundaries.
• Validation was performed on a dataset from the Signal Processing laboratory, Brno university.

Conclusions:

• The proposed automated method provides highly accurate and reliable segmentation of the lumen intima layer.
• This technique has the potential to enhance the quantitative analysis of carotid artery ultrasound images.
• The hybrid approach effectively combines coarse and fine segmentation strategies for improved performance.