粒子群优化算法选择特征的运动图像分类  

Moving image classification based on particle swarm optimization algorithm selecting features

作　　者：吴雪[1] WU Xue(Wuhan Institute of Technology,Wuhan 430205,China)

机构地区：[1]武汉工程大学,湖北武汉430205

出　　处：《现代电子技术》2017年第17期47-50,共4页Modern Electronics Technique

摘　　要：为了提高图像分类的效果,考虑当前方法准确实现图像分类的难题,提出粒子群优化算法选择特征的运动图像分类方法。对当前运动图像分类方法的研究现状进行分析,提取不同类型的图像,并采用粒子群优化算法选择最优特征,组成特征向量,将特征向量机作为神经网络的输入,实现运动图像的分类。采用具体图像分类实验进行验证,结果表明,该方法可以描述不同运动图像的类别信息,缩小图像的分类误差,避免其他图像分类方法的缺陷,提高了图像的整体分类正确率。In order to improve the effect of image classification and realize the accurate image classification,a moving imageclassification method based on particle swarm optimization algorithm selecting features is proposed.The current research statusof the moving image classification methods is analyzed to extract the images of different types.The particle swarm optimizationalgorithm is used to select the optimal feature to compose the feature vector.The feature vector machine is taken as the input ofneural network to classify the moving images.The classification experiments of specific images were adopted to make verification.The experimental results show that the method can describe the categories information of different moving images,reduce theclassification error of the images,avoid the defects of other image classification methods,and improve the overall image classification accuracy.

关 键 词：运动图像 特征选择 粒子群算法 图像分类 

分 类 号：TN911.73-34[电子电信—通信与信息系统] TP317.4[电子电信—信息与通信工程]