中国计量大学学报2017，Vol.28Issue(2)：234-240,268,8.DOI:10.3969/j.issn.2096-2835.2017.02.016
特征融合的多视角步态识别研究
Research on multi-perspective gait recognition using feature fusion
摘要
Abstract
Since the low gait recognition rate of single features, a novel gait recognition algorithm based on the feature fusion of the dynamic part of gait energy images (GEI) and Gabor was presented in this paper.Firstly, the gait contour images were extracted through the object detection, binarization and morphological process, to caluclate GEI and divide the dynamic part of GEI.Secondly, the information of different angles was extracted from the dynamic part of GEI with Gabor wavelets.Feature fusion was extracted with GEI and Gabor wavelets and was reduced by improved KPCA.Finally, the vectors of feature fusion were input into the SVM (Support Vector Machine) based on multi classification to realize the classification and recognition of gait.Experiments were conducted on the Central Asia Student International Academic(CASIA) gait database with satisfactory recognition effect.Compared with methods based on the single gait feature, the gait recognition rate after the fusion was executed increased about 10%.关键词
Gabor小波/步态能量图/特征融合/改进的KPCA/支持向量机Key words
Gabor wavelets/GEI/feature fusion/improved KPCA/SVM分类
信息技术与安全科学引用本文复制引用
王竣,王修晖..特征融合的多视角步态识别研究[J].中国计量大学学报,2017,28(2):234-240,268,8.基金项目
国家自然科学基金资助项目(No.61303146). （No.61303146）
