基于C-V模型的眼底图像交互式杯盘分割

INTERACTIVE SEGEMENTATION FOR OPTIC CUP-DISK OF FUNDUS IMAGE BASED ON C-V MODEL

针对眼底图像视杯和视盘水平集分割中C-V模型自适应能力不强等问题,提出一种基于C-V模型的视盘和视杯交互式水平集分割算法。该方法通过交互方式给定不同的视盘初始轮廓和C-V模型参数,对眼底图像的杯盘进行精确地分割。实验结果表明,该方法可克服噪声污染、光照不均匀、对比度低等特点对眼底图像分割的影响,对彩色眼底图像中的视杯和视盘进行精确分割。

C-V model is weak at self-adaptation in optic cup-disk of fundus image and level set segmentation.So the authors propose a C-V model based optic disk and optic cup interactive level set segmentation method.The method,through an interactive way,provides different optic disk initial contour and C-V model parameters to precisely segment fundus image cup-disk.Experimental results show that the method can overcome impacts from noise pollutions,uneven illumination,obscure contrast as well as other shortcomings on fundus image segmentation so as to precisely segment optic cup-disk in a color fundus image.