摘要
提出一种无须重新初始化的变分水平集自适应主动轮廓模型。该模型利用图像的局部特性自适应决定曲线的演化,同时加入局部C-V能量项,改进边界停止函数,提高对灰度分布重叠、分布不均匀及弱边界处理的鲁棒性,并加快了曲线演化的收敛速度。结合医学序列图像特点,利用Heaviside函数对当前截面分割结果进行分段常量化后投射至相邻界面作为初始化曲线,实现对序列图像的自动分割。最后,以骨关节磁共振图像中正常结构和病灶组织的分割实验对算法进行了验证。
Art adaptive active contour model-variational level set without re-initialization is proposed in this paper. We use local characteristics of image to implement adaptive curve evolution. Simultaneously, We add local C-V energy term and improve the edge stop function. These can increase the iterative convergence speed and make it more robust to the intensity distribution overlapping, intensity inhomogeneity and weak boundary. Combining with the characteristics of serial medical images, transform the segmentation result of current slice to piecewise constant function using the Heaviside function. Cast it to the next slice as the initial curve to implement automatic segmentation of serial images. The proposed method has been applied to segment normal structure and diseased tissue in musculoskeletal MRI with promising results.
出处
《中国图象图形学报》
CSCD
北大核心
2011年第7期1199-1205,共7页
Journal of Image and Graphics
基金
安徽省2010高校省级自然科学研究重点项目(KJ2010A193)
教育部博士点基金项目(20060359004)
教育部留学归国人员科研启动基金项目(413117)