遗传神经网络在图像分割中应用研究  

A Study of Genetic Neural Network Used in Image Segmentation

作　　者：丁红军[1] 蔡鸿杰[1] 邢克礼[2] 

机构地区：[1]天津公安警官职业学院侦查系,天津300382 [2]天津医科大学生物医学工程系,天津300070

出　　处：《自动化技术与应用》2010年第3期8-12,共5页Techniques of Automation and Applications

摘　　要：针对图像分割的复杂性,利用遗传算法对BP神经网络的权值和阈值进行优化,设计出误差最小的神经网络,然后再对图像的像素进行分类识别,实现并提高了图像分割性能。仿真实验表明,与传统的图像分割方法相比,取得了比传统方法更好的图像分割效果。Because of the complexity of image segmentation,the optimization of the weights and thresholds of BP neural network are realized by genetic algorithm,and a BP neural network with minimum error is designed,It classify the image pixels,implement and improve the performance of image segmentation.The results of simulation show that the algorithm neural network can better achieve the image segmentation,compared with the traditional method.

关 键 词：图像分割 神经网络 遗传算法 遗传优化 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]