摘要
提出了一个基于径向基函数网络的医学图像分类器。该系统包括图像预处理、特征提取、分类器的构造几个部分。在网络构造上,文章采用了自适应的网络结构调整技术,提高了网络的泛化能力,同时,在网络权值调节上采用了具有全局优化的模拟退火算法,避免了陷入局部极小的缺陷。实验结果表明,该模型系统达到了76.6%的准确率,辅助系统可以极大地提高医学图像分类的效率和准确性。
This paper proposes a medical images classifier based on radial basis function neural network.The system we propose consists of a pre-processing phase,a feature extraction phase and a building the classifier phase.The system uses adaptive network structure adjustment techniques to improve the generalization capability of the network.At the same time ,to keep the network from getting into local minimum the system uses simulated annealing with feature of global optimize to adjust the network weight.The experimental results show that the system performs well reaching about 76.6%in accuracy.With the help of the system physicians can improve efficiency and accuracy in the medical image classification.
出处
《计算机工程与应用》
CSCD
北大核心
2004年第17期230-232,共3页
Computer Engineering and Applications
基金
国家自然科学基金项目(编号:60271032)资助
关键词
径向基函数网络
图像分类
模拟退火算法
医学图像
乳腺X线照片
radial basis function networks,images classification,simulated annealing,medical images,mammography