基于 KNN 的特征自适应加权自然图像分类研究

针对自然图像类型广泛、结构复杂、分类精度不高的实际问题,提出了一种为自然图像不同特征自动加权值的K-近邻(K.nearestneighbors,KNN)分类方法。通过分析自然图像的不同特征对于分类结果的影响,采用基因遗传算法求得一组最优分类权值向量解,利用该最优权值对自然图像纹理和颜色两个特征分别进行加权,最后用自适应加权K-近邻算法实现对自然图像的分类。实验结果表明,在用户给定分类精度需求和低时间复杂度的约束下,算法能快速、高精度地进行自然图像分类。提出的自适应加权K-近邻分类方法对于门类繁多的自然图像具有普遍适用性,可以有效地提高自然图像的分类性能。...

Full description

Saved in:
Bibliographic Details
Published in计算机应用研究 Vol. 31; no. 3; pp. 957 - 960
Main Author 侯玉婷 彭进业 郝露微 王瑞
Format Journal Article
LanguageChinese
Published School of Information & Technology,Northwest University,Xi'an 710127,Chin 2014
Subjects
Online AccessGet full text
ISSN1001-3695
DOI10.3969/j.issn.1001-3695.2014.03.077

Cover

More Information
Summary:针对自然图像类型广泛、结构复杂、分类精度不高的实际问题,提出了一种为自然图像不同特征自动加权值的K-近邻(K.nearestneighbors,KNN)分类方法。通过分析自然图像的不同特征对于分类结果的影响,采用基因遗传算法求得一组最优分类权值向量解,利用该最优权值对自然图像纹理和颜色两个特征分别进行加权,最后用自适应加权K-近邻算法实现对自然图像的分类。实验结果表明,在用户给定分类精度需求和低时间复杂度的约束下,算法能快速、高精度地进行自然图像分类。提出的自适应加权K-近邻分类方法对于门类繁多的自然图像具有普遍适用性,可以有效地提高自然图像的分类性能。
Bibliography:K-nearest neighbors (KNN) algorithm ; genetic algorithm ; natural images classification ; feature-weighted
In order to solve the natural images problems of widely types, complex instruction and low classification accuracy, this paper proposed a new classification method for natural images based on feature-weighted K-nearest neighbors. By analyzing the impact of different features on natural images classification, using genetic algorithm to get a set of optimal classification weight vector, weighted textural features and color features based on that data. Finally, it used the adaptive feature-weighted K-nearest neighbors to classify the natural images. The experimental results show that in the constraint Of demand classification accuracy and low time complexity that the user given, this algorithm can classify the natural images with high speed and high- precision. The adaptive feature-weighted K-nearest neighbors classification method is generally apply for the variety of natural images, and can improve the cla
ISSN:1001-3695
DOI:10.3969/j.issn.1001-3695.2014.03.077