Multi- View Gait Human Image Generation for Cross-View Gait Recognition by Back Propagation Neural Networks

Abstract

Human gait recognition to recognize persons identities by walking styles, pedestrian walking
speed, angle changing, etc. Gait recognition has idiosyncratic advantages due to its
characteristics of non-contact and long distance compared with face and fingerprint
recognition. Because view variance gait silhouettes, Cross- view gait detection is a difficult
task. In the existing system, for all view in a single or many data sets, it can improve
generalization capacities of gait classification models. The MVGGAN technique trained a
single generator. They also performed domain alignment based on predicted maximum mean
discrepancy to limit the impact of sample generation-induced distribution divergence. We
proposed the new method for recognizing humans by their gait using BPNN. The BPNN
algorithm is used to recognize humans by their gait patterns. Automatic gait recognition using
Fourier descriptors and ICA for the purpose of human identification at a distance. Firstly, a
simple background generation algorithm is introduced to subtract the moving figures accurately
and to obtain binary human silhouettes. Secondly, these silhouettes are described with Fourier
descriptors and converted into associated one-dimensional signals. Then ICA is applied to get
the independent components of the signals. For reducing the computational cost, a fast and
robust fixed-point algorithm for calculating. Gait recognition aims essentially to address this
problem by identifying people based on the way they walk Gait recognition has three steps.
The first step is pre-processing, the second step is feature extraction and the third one is
classification. This project focuses on the classification step that is essential to increase the
CCR. The MLP is used in this work. In this project we apply MLP NN for eleven views in our
database and compare the CCR values for these views. In experiments, higher gait recognition
performances have been achieved.