Gait recognition based on feature fusion and support vector machine

Facing the problem that the recognition rate based on single feature algorithm is low in present gait recognition studying, a new gait recognition algorithm based on feature fusion and support vector machine is proposed. First, getting the body contour image by using background subtraction, then ext...

Full description

Saved in:
Bibliographic Details
Published in2016 IEEE International Conference of Online Analysis and Computing Science (ICOACS) pp. 281 - 284
Main Authors Ye Hanmin, Huang Peiliang
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.05.2016
Subjects
Decision support systems
dynamic regional variance
feature fusion
Fourier descriptor
gait recognition
limb angle
Nickel
support vector machine
Support vector machines
Target recognition
Online AccessGet full text

Cover

More Information
Summary:Facing the problem that the recognition rate based on single feature algorithm is low in present gait recognition studying, a new gait recognition algorithm based on feature fusion and support vector machine is proposed. First, getting the body contour image by using background subtraction, then extracting the human body contour line and using Fourier descriptor to express human body contour feature, extracting limb angles feature, also introducing a kind of dynamic regional variance feature. Fusing them to get a combination features vector and finally using support vector machine to classify the vector. A plenty of experiments are made on the CASIA gait databases. Experimental results show the high recognition rate of this method.
DOI:10.1109/ICOACS.2016.7563097