Performance Analysis for Human Gait Recognition from different Viewpoints

Human gait, a new biometric aimed at recognizing people based on how they walk, has become increasingly important in visual surveillance application. However, one camera single view point gait data commonly been explored, this not always sufficient enough the environment of deployment. This research proposes analysis as solution for subjects’ identification across network cameras from different viewpoints. Gait signature person is created Temporal and spatial metrics extracted modal, such length trunk, shin deviation limb angles or amplitude person’s walking pattern these are transformed into self-similarity matrix. The method spatio-temporal correlation to detect human successive video sequences. Performance evaluation system was carried out with CMU (Carnegie Mellon University) Motion Body (MoBo) database. results reveal that performance proposed possible even without knowing position stance subject. shows derived parameters suggest may be successfully employed individuals' scenario which resulted an average recognition rate 73.3% persons right viewpoints, 80.0% left viewpoint, 60.0% rare viewpoint 66.67% front viewpoint. implies performs better