基于SVM算法的两种特征提取的图像分类方法分析  

Analysis of Image Classification Methods Based on Two Feature Extraction Based on SVM Algorithm

作　　者：王瑶 徐昌 舒福舟 WANG Yao;XU Chang;SHU Fu-Zhou(School of electronic information,Yangtze University,Jingzhou 434023,China)

机构地区：[1]长江大学电子信息学院

出　　处：《电脑与信息技术》2019年第6期18-20,33,共4页Computer and Information Technology

摘　　要：为了提高图片分类的效果,文章提出了基于SVM算法的两种特征提取的图像分类的方法。首先分别对图片数据集进行灰度直方图特征提取和SIFT特征提取,然后再用SVM算法对以上两种特征提取的图片数据集进行分类。实验结果表明,采用SIFT特征提取的图片分类效果明显好于采用灰度直方图的图片分类效果。In order to improve the effect of image classification,this paper proposes a method of image classification based on SVM algorithm.Firstly,the grayscale histogram feature extraction and SIFT feature extraction are performed on the image dataset respectively,and then the SVM algorithm is used to classify the image datasets extracted by the above two features.The experimental results show that the image classification effect using SIFT feature extraction is significantly better than that of grayscale histogram.

关 键 词：灰度直方图 SIFT特征提取 SVM 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]