模糊B样条基神经网络磁共振图像分割方法

A Segmentation Method of MRI Based on Fuzzy B-spline Basis Neural Network

针对磁共振图像分割的特点,提出了一种基于模糊B样条基神经网络的磁共振图像分割方法。该方法采用B样条基函数作为模糊隶属函数,利用神经网络实现模糊推理,并采用反向误差传播算法对网络进行训练。实验结果表明,这种基于模糊B样条基神经网络的磁共振图像分割方法与普通神经网络分割方法相比,具有更高的分割精度和更快的训练收敛速度。

Considering the features of magnetic resonance image（MRI） segmentation,a segmentation method of MRI based on fuzzy B-spline basis neural network is proposed.This method uses B-spline basis function as fuzzy membership functlon,neural network to realize fuzzy inference,and BP algorithm to train network.The results of experiments show that the proposed method has better segmentation precision and faster network convergence speed than traditional neural network segmentation method.