研究目的
To develop an Elliptical Active Disc (EAD) methodology to delineate the Lumen Intima Layer (LIL) of the Common Carotid Artery (CCA) from transverse mode ultrasound images for accurate, inexpensive, and noninvasive progression and regression monitoring of atherosclerosis.
研究成果
The proposed Elliptical Active Disc method achieves a CCA detection accuracy of 97.63% and an average Dice similarity index of 94.83% on SPLab ultrasound images. It offers competitive detection and segmentation accuracy over previously proposed methods and is available as an ImageJ plugin. Future research could focus on automatic carotid plaque quantification and stenosis analysis using the proposed segmentation technique.
研究不足
The study is focused on transverse mode ultrasound images of the common carotid artery. The method's performance on other types of images or arteries is not evaluated. The computational efficiency and accuracy might vary with image quality and resolution.
1：Experimental Design and Method Selection:
The methodology involves the optimization of a local energy function with respect to five degrees-of-freedom that characterize the elliptical active disc. Gradient descent technique is used for optimization. Green's theorem is applied to optimize the computation of partial derivatives.
2：Sample Selection and Data Sources:
The experiments are carried out on a database provided by SPLab, Brno University, containing 971 transverse mode ultrasound images of the carotid artery.
3：List of Experimental Equipment and Materials:
A desktop with MacOS X
4：7 GHz, Intel Core i5 processor is used. Color images are converted to inverted grayscale and histogram-equalized for processing. Experimental Procedures and Operational Workflow:
Automatic initialization of the EAD is performed using the normalized cross-correlation technique. The EAD evolves based on a predefined energy criterion to perform segmentation.
5：Data Analysis Methods:
The similarity between the ground truth and algorithm segmentation is quantified using the Dice index.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
