location: Current position: Home >> Scientific Research >> Paper Publications

低分辨率多姿态人脸识别算法研究

Hits:

Indexed by:期刊论文

Date of Publication:2016-07-20

Journal:控制工程

Included Journals:ISTIC、CSCD

Volume:23

Issue:7

Page Number:1057-1062

ISSN No.:1671-7848

Key Words:深度信念网络;极限学习;低分辨率;姿态变化;人脸识别

Abstract:人脸识别一直是模式识别和机器视觉领域研究热点.受姿态变化和分辨率低影响,传统方法对伴随多姿态低分辨率人脸图像识别精度较低.为充分挖掘姿态变化带来的非线性问题,提出将深度信念网络与极限学习机相结合来识别低分辨率多姿态人脸图像.该方法将低分辨率和对应高分辨率图像作为深层网络结构输入数据,学习高低分辨率图像间流行假设的点对联系以提取特征进行分类识别.实验结果表明,所提方法相比于其他方法具有识别率高、分类时间短等优点.

Pre One:Predicting pupylation sites in prokaryotic proteins using semi-supervised self-training support vector machine algorithm

Next One:Predicting lysine phosphoglycerylation with fuzzy SVM by incorporating k-spaced amino acid pairs into Chou's general PseAAC