摘要
朴素贝叶斯分类器是分类算法集合中基于贝叶斯理论的一种算法,为了对海量的视频进行分类,提出了一种基于朴素贝叶斯分类器的视频分类方法。首先,对视频进行特征提取,提取纹理、颜色以及亮度三种反映其类别的特征。在提取到视频的颜色、亮度以及纹理特征之后,然后进行基于朴素贝叶斯分类器的视频分类。对于输入的几种特征,采用极大似然估计,直至完成最终的分类。实验结果表明,提出的视频分类方法取得了较为准确的分类结果,对篮球、足球、斯诺克等5个运动项目均取得了90%以上的分类准确性。
Naive Bayesian classifier is an algorithm based on Bayesian theory in the collection of classification algorithms.In order to classify a huge number of videos,a video classification method based on naive Bayesian classifier was proposed.First,feature extraction is performed on videos,and three features,namely texture,color,and brightness,which reflect its category,are extracted.After the color,brightness,and texture features of the video are extracted,video classification based on the naive Bayes classifier is then performed.For the input features,maximum likelihood estimation is used until the final classification is completed.The experimental results show that the classification method of videos obtained has obtained more accurate classification results,and the classification accuracy of more than 90% has been achieved for five such as basketball,football,and snooker.
作者
庞博
成东坡
Pang Bo;Cheng Dongpo(Shangqiu polytechnic,Shangqiu 476100,Henan;Jiyuan High school,Shangqiu 459000,Henan)
出处
《武汉工程职业技术学院学报》
2019年第3期14-17,共4页
Journal of Wuhan Engineering Institute
关键词
视频分类
朴素贝叶斯
特征提取
极大似然估计
video classification
naive Bayes
feature extraction
maximum likelihood estimation
