Gait based gender classification using Kinect sensor
Authors
Chen, Yan
Yang, Yawei
Lee, Jeongkyu
Issue Date
2014-03-28
Type
Presentation
Language
en_US
Keywords
Engineering , Faculty research day , Kinect sensor , Gait analysis
Alternative Title
Abstract
In this project, we propose a novel method to recognize human gender based on their gaits. We collect samples of walking silhouettes with Microsoft Kinect sensor and extract gait features from Gait Energy Image (GEI). The samples are divided into two parts: training dataset and testing dataset. We train a SVM classifier using the training set and test with testing dataset. We use feature vector with a low dimension in this project. The experimental results show that our method has accuracy higher than 80%.