SGNN优化算法的研究及其在图像分割中的应用  

Optimized Self-Generating Neural Network for Image Segmentation

作　　者：李露[1] 

机构地区：[1]北京航空航天大学,宇航学院,北京100191

出　　处：《红外技术》2010年第4期198-203,共6页Infrared Technology

摘　　要：针对传统神经网络用于图像分割中存在着网络结构设计复杂、计算量大等缺点,提出了一种基于自生成神经网络(Self-Generating Neural Network,SGNN)的图像分割方法,将图像的每个像素按其灰度值自动聚类,从而实现图像的自动分割。在此基础上,本文着重研究了SGNN网络的优化算法,以期达到更好的分类效果。实验结果表明,该方法可以很好的实现图像分割,无需人为干涉,具有学习自主性高,分类效果明显,抗噪能力强等优点,可广泛用于红外、可见光、X光、MR等多种图像的分割。In this paper, we propose a novel SGNN-based image segmentation method, in which, not only the network structure, but also the weights among neurons are all acquired automatically through the training samples. Therefore, image segmentation is implemented automatically by autonomously clustering the pixels according to their gray values. Then the optimization of SGNN is studied to get better segmentation results. The experimental results show that the optimized SGNN outperforms the existing methods for its distinguished advantages of perfect segmentation without any manual intervention, high self-learning capacity, less computational complexity, robustness to noise, etc. what＇s more, the experimental results suggest the proposed method can be widely used in segmentation of all typical images, such as IR （Infrared） images, visible light images, X-ray images, MR （Magnetic Resonance） Images.

关 键 词：神经网络 SGNN 图像分割 优化 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]