基于模糊集与区域生长算法的胎儿下腔静脉血管超声分割

本方研究了胎儿下腔静脉血管在B型超声图像中的分割问题.B型超声使用方便，在临床中有广泛使用，但其图像有噪声多、对比度差的缺陷.为了有效地在B型超声图像中分割血管，提出了一种基于模糊集与区域生长算法的分割算法；该算法预先使用模糊集算法处理，以提高图像对比度；并使用基于梯度改进的自适应区域生长算法进行分割.实验以医生的手工分割结果作为金标准，并与阈值分割和水平集算法进行了对比.实验表明，该方法的准确度和稳定性高于阈值分割和水平集分割方法结果.

A Segmentation Algorithm Based on Fuzzy Sets and Region Growth

A segmentation method was studied for the inferior vena cava of fetus in B-mode ultrasound images in this paper. B-mode ultrasound is a technique widely used in clinic for its convenience, but the images of that have the defect of high noise and low contrast. To effectively segment blood vessels in B-mode ultrasound images, in this paper, a segmentation algorithm based on fuzzy sets and region growth algorithm was presented. A fuzzy set algorithm was used to improve the image contrast in advance, and then the adaptive region growth algorithm was presented based on gradient to segment the image. In this paper, taking the doctor''s manual segmentation result as the gold standard, the threshold segmentation and level set algorithm were compared. The experiments show that, the accuracy and stability of this method are higher than threshold segmentation and level segmentation.