摘要
以龋齿诊断为例,探讨了灰度共生矩阵和神经网络在医学图像处理中的应用。通过对患者龋齿图像的特征分析,采用从灰度共生矩阵中提取的4个参数作为神经网络的输入特征向量,经过对该神经网络的多次训练,实现龋齿的识别。利用Matlab与VC++语言来设计龋齿诊断程序,并借助MIDEVA将其转化为脱离Matlab的工作环境的可执行程序,大大节省了系统资源。
The application of gray level co-occurrence matrix and neural network in medical image processing by taking tooth decay diagnosis as an example was proposed and carefully experimented.With four coefficients extracted form gray level co-occurrence matrix of the tooth images as its input feature vector by analysis,the network was used to make differential diagnoses between decayed and normal teeth after it has been trained for many times.The diagnosis programs were projected with Matlab and VC++,and they were transformed into executable programs independent of Matlab,which save efficiently the limited system hardware resources.
出处
《实验技术与管理》
CAS
北大核心
2011年第7期59-61,共3页
Experimental Technology and Management
关键词
龋齿识别
灰度共生矩阵
神经网络
MIDEVA
tooth decay diagnosis
gray level co-occurrence matrix
neural network
MIDEVA