Research On Face Recognition System Based On Sparse Representation And Machine Learning

Face recognition is one of the most challenging topics in the field of machine vision and pattern recognition,and also has a wide range of applications.In e-commerce,public security systems,file management,human-computer interaction and other fields have a high application and commercial value.However,the current face recognition system still has many defects in practical application,such as the illumination,the environment and the attitude,the recognition rate is not satisfactory.Therefore,this thesis attempts to find a combination of machine learning and sparse representation of the face recognition system,it can be in a noisy environment with better recognition rate.The main work and contributions of this thesis are:(1)Firstly,the problem of face recognition is not good in the noise environment,and a recognition algorithm based on sparse representation is used to improve the recognition rate under occlusion or poor lighting condition.(2)Due to the lack of the traditional face feature extraction method,the extracted feature is too simple and elementary.To solve this problem,a cascade convolution self-coder based on depth learning theory is used to extract features.Cascaded convolutional self-coder improves the recognition rate of the system by self-learning,reducing the error of feature extraction,and extracting more deep and abstract features by multi-level cascade.(3)Two kinds of methods are put forward: the sparse representation based classifier algorithm and the stacked convolution auto-encoder based on deep learning theory.A face recognition system is designed and implemented.This system has a good performance in the noise environment,and the recognition speed is fast and the accuracy is high.