计算机技术与发展Issue(2)：31-34,4.DOI:10.3969/j.issn.1673-629X.2016.02.007
基于PCANet-RF的人脸检测系统
Face Detection System Based on PCANet-RF
摘要
Abstract
A face detection system was presented based on a simple convolutional neural network. Feature extraction of image is usually complicated which needs much pretreatment. Deep learning reduces pretreatment,such as convolutional neural network,but it needs more time of training and requires certain ability to adjust the parameters,which contrary to the original intention. What is more,classification capability and result of convolutional neural network is not well. Combination of above,the PCANet for feature extraction is applied to lower the ability to adjust the parameters and Random Forest for image classification is used to improve the recognition rate. This method has got a recognition rate as 99%. Experiments has confirmed that PCANet-RF can be successfully used in image classification.关键词
人脸检测/卷积神经网络/随机森林/特征提取/主成分分析网络Key words
face detection/convolutional neural network/random forest/feature extraction/PCANet分类
信息技术与安全科学引用本文复制引用
张丹丹,李雷..基于PCANet-RF的人脸检测系统[J].计算机技术与发展,2016,(2):31-34,4.基金项目
国家自然科学基金资助项目(61070234,61071167,61373137) （61070234,61071167,61373137）
江苏省普通高校专业学位研究生科研实践计划省立(SJLX 0376) （SJLX 0376）
