Research on Gender Classification Based on Gait Features
This paper designs a set of gender classification model based on gait features.The input of the model is in the form of an energy graph.The model adopts a series-linked network structure and the Softmax function for the final character gender classification.Then,the gait profile in the CASIA-B database is clipped to synthesize an energy profile and input into the model for gender classification testing.The overall classification accuracy is more than 91％.Finally,an experiment is designed to compare the proposed model with the VGG and ResNet classification models.The experimental re-sults show that the model has better generalization performance and efficiency for character gender classification.
万方数据
维普
