基于可变区域拟合水平集算法的改进

An improved image segmentation algorithm based on RSF

基于可变区域拟合水平集算法利用图像的局部区域信息,在活动曲线演化控制参数的手工设置使其应用受到了限制。本文提出了将灰度信息图像匹配原理应用到RSF模型中,根据计算相邻演化图像的相关系数实现迭代的自适应停止。实验结果表明,改进的RSF模型克服了自动设置迭代次数的缺点,实现了迭代的自适应停止,而且对弱边缘不连续图像能够有效地实现,节省了时间,提高了分割效率。

Base on Region-scalable fitting level set algorithm using local area information,but the curve evolution control parameters manually set way make its application is limited.Information is proposed in this paper the gray scale image matching principle is applied to the RSF model,according to the calculation of adjacent image correlation coefficient to realize iterative adaptive evolution.The experiental results show that the improved RSF model overcomes the drawback of automatically set the number of iterations, implements the iterative adaptive to stop,and the weak edge discontinuity image effectively,save time and improve the efficiency of segmentation.