基于细胞神经网络的数学形态滤波方法与应用

CNN based mathematical morphological filter with applications

数学形态学以集合运算为基础,在图像处理领域得到了广泛地运用。数学形态学以集合运算为基础,用具有一定形态的结构元素去度量图像中的形态以解决理解问题。证明利用细胞神经网络(Cellular Neural Network),运用数学形态滤波可并行完成数学形态运算。该文给出了细胞神经网络(CNN)在腐蚀、膨胀、结构开、结构闭中的实现及应用。将其结果运用在指纹图像的预处理当中,取得了较理想的结果。

Mathematical morphology is widely applied in digital image proeessing.The base idea of mathematical morphology is to use construction element measure image morphology for solving understand problem.The paper presents Cellular Neural Network （CNN）to be suit for mathematical morphology filter.It gives realizations and applications for erosion,dilation,open and close operations based on CNNs.An example for fingerprint image processing illustrates the effectivities of the approach.