基于BP算法的纹理图象合成研究  

Study on Texture-image Synthesis Based on BP Algorithm

作　　者：彭晏飞[1] 刘树启[1] 

机构地区：[1]辽宁工程技术大学电子与信息工程学院,葫芦岛125105

出　　处：《世界科技研究与发展》2009年第3期462-464,共3页World Sci-Tech R&D

摘　　要：针对目前纹理图象合成技术存在的弱点,应用人工神经网络BP算法实现了纹理图象的合成。通过对纹理图象的分析和特征提取来获取图象的各种参数信息,然后采用人工神经网络BP算法,建立BP网络模型,将原始图象的特征函数作为BP网络的输入,通过学习-训练,确定该图象的人工神经网络结构,并最终通过程序合成了纹理图象。To solve the present problems in texture-image synthesis,applying the Artificial Neural Network BP algorithm so that the texture- image synthesis is accomplished smoothly. The texture-image analysis and feature extraction is used to obtain images of a variety of parameters, and then using the Artificial Neural Network BP algorithm, a BP network model is established the original image as a function of the characteristics of the BP network input, through the Study-Training to determine the image of the artificial neural network structure, and eventually through the process of synthesis of texture-images.

关 键 词：纹理图象合成 BP算法 特征函数 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]