引入新的停止函数的GAC模型医学图像序列分割

摘要

针对GAC模型在低对比度及边缘模糊图像分割中使用传统的边缘停止函数存在的不足,结合图像的区域特征及医学图像序列间的相关特性,通过在边缘停止函数中引入图像的区域特征信息和先验信息,对GAC模型的边缘停止函数进行改进,并应用于医学CT图像序列的分割,达到了较为理想的分割结果.实验表明,在边缘停止函数中引入先验信息并对其表达形式进行改进后,模型的抗噪声性能和分割效率得到了明显提高.%Traditional geodesies active contour model has a shortage in soft image and edge fog image segmentation. This paper constructed a new edge stop function by combining regional characteristics with relevance between sequence image, and brought the regional characteristics and priori information into the edge stop function. Then it applicated this new function in GAC model to improve medical sequence image segmentation quality, and achieved a much ideal result. Experiments show that after introducting of a priori information and improving its forms of expression, it greatly improved anti-noise performance and segmentation efficiency of the model.