基于离散水平集方法的Chan-Vese模型

采用离散形式的水平集函数代替传统的符号距离函数,提出基于离散水平集方法的Chan-Vese模型。采用变分方法得到水平集函数演化的梯度降方程,对得到的演化方程进行半隐式离散求解和数值算例验证,并与传统的符号距离函数方法进行对比,以验证离散水平集方法在图像分割中的优势。实验结果表明,基于离散水平集方法的Chan-Vese模型分割速度和精度比用符号距离函数实现的Chan-Vese模型有明显提高。

Chan-Vese Model Based on Discrete Level Set Method

A new method which uses discrete definition of the level set function instead of signed distance function definition to compute the Chan-Vese model is proposed.The gradient flow of the functional is computed using variational method to get the optimized result of the model and is discredited by semi-implicit difference method.Comparison with the signed distance function is conducted to verify the advantages of the discrete level set method in image segmentation.Experimental results shows that proposed model can segment image in a faster and more accurate way than using the signed distance function.