自组织神经网络在图像处理中的应用  

The Application of Self-organising Neural Network in Image Processing

作　　者：丰艳[1] 陈一民[1] 吴志扬[2] 

机构地区：[1]上海大学计算机工程与科学学院,上海200072 [2]上海工业自动化仪表研究所,上海200233

出　　处：《自动化仪表》2005年第8期32-34,共3页Process Automation Instrumentation

摘　　要：针对医学图像以及医学观察的特点,提出利用Kohonen神经网络进行交互式的医学图像分割。针对常见的聚类数目确定方法在图像处理中存在的不足,本文提出神经网络中输出结点的个数(即图像分割的类别数)不是固定按照某一种或某几种以往常见的方法来确定,而是由医生根据分析病情的需要而动态决定,增加了灵活性和医生的自主性。另外,针对图像处理数据量大、处理速度慢的特点,利用PDS塔型数据结构提高图像分割的速度。In accordance with the characteristics of medical images and medical observation, the method of interactive medical image segmentation by using Kohonen neural network is stated. To against the disadvantages of contain methods for determining the number of clusters in image processing, a new concept is proposed. That means the number of output nodes of neural network （ the number of categories in image segmentation） is not fixed determined by one or several common methods, but it is detemtined dynamically by physician＇s analysis on patients＇ conditions. The concept enhanced flexibility and doctors＇ independence. In addition, the speed of segmenting images is enhanced by using PDS data structure to meet the features of image processing, i.e., large amount of data and slow in processing.

关 键 词：KOHONEN神经网络 图像分割 输出结点 交互式 PDS 自组织神经网络 图像处理 医学图像分割 应用 处理速度 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] TP391.41[自动化与计算机技术—控制科学与工程]