重庆理工大学学报（自然科学版）2017，Vol.31Issue(10)：192-197,6.DOI:10.3969/j.issn.1674-8425(z).2017.10.031
训练样本数量选择对图像特征提取的影响分析
Influence of the Number of Training Sample on Image Feature Extraction
摘要
Abstract
Image feature extraction is one of the important contents of image processing.The quality of feature extraction directly affects the effect of image classification,image recognition and image retrieval.There are many factors that affect image feature extraction.The restricted Bohzmann machine is took as an example to discuss the influence of the number of training samples on image feature extraction.The experimental results show that the generality of the image features extracted by the restricted Boltzmann machine with the same parameters will be enhanced as the number of samples increases.Therefore,when the number of samples is large,increasing the number of hidden layer neurons in RBM is not always valuable.关键词
图像特征提取/受限玻尔兹曼机/CD算法Key words
image feature extraction/restricted Boltzmann machine/CD algorithm分类
信息技术与安全科学引用本文复制引用
尹静,闫河..训练样本数量选择对图像特征提取的影响分析[J].重庆理工大学学报（自然科学版）,2017,31(10):192-197,6.基金项目
国家自然科学基金资助项目(61173184) （61173184）
