带评价系统的曲率相关有向主动轮廓模型

主动轮廓模型广泛应用于很多领域,但这个框架下的模型都有两个关键的难点.针对这一问题,通过对已有方法失败原因的分析,本文提出了一种新颖的外力大小与曲率相关的主动轮廓模型（CDM模型）.任何粗分割方法都能为该模型预测方向提供灵活有效的线索,并且模型外力的大小定义为曲线曲率的相关方程.这使得该模型具有许多与众不同的优点,并能够有效避免自相交现象的发生.在不同的收敛过程和不同的主动轮廓模型中,评价模块的出现能够有效地度量模型结果与真实目标轮廓之间的相似度.实验结果显示,CDM模型在其他模型失败的例子中能够得到正确、鲁棒的结果,并且能够在医学图像中精确地拟合目标轮廓.

Directional Active Contour Model Based on Curvature Dependent Magnitude with Evaluation System

Active contour models are widely used in many applications,however,most models under active contour framework have two key difficulties.Aiming at this,this paper presents a novel method called curvature-dependent magnitude model（CDM model）.Any crude segmentation methods are feasible to provide cues for the CDM model to forecast the moving direction,and the external force magnitude is a related function of curve curvature which makes the model has several advantages as well as avoiding the self-intersection problem.The evaluation system is valid to measure the approximation degree between the result and the real boundary in different convergence cases and different models.The experimental results show that the CDM model is robust and efficient in failure cases of the other models and able to gain good performance on fitting the object boundary in medical images.