Research On Face Recognition Based On Wavelet Neural Network

With the continuous development of information technology and the gradual advancement of the era of big data,face recognition technology has become a hot research field combining with popular subjects such a machine learning,pattern recognition and deep learning,etc.And it has won the focus of researchers from all walks of life in recent years.There are many scenes using face recognition in our life at present,and the recognition algorithm is too numerous to mention.But the algorithm represented by neural network has been applied with excellent performance widely and it has achieved remarkable results among them.As a kind of neural network,wavelet neural network has been recognized for its good basis of wavelet analysis by the scientific research circle.This thesis focuses on the study of face recognition algorithm based on wavelet neural network.We integrate and innovate from the aspects of feature extraction,identification,ensemble learning and parameter optimization,etc.These can achieve the better ability of learning and expression,and the final results can be more accurate and stable.First of all,we give a brief overview for the research topic of the background and so on in the first chapter.The development,application and research status of neural network and face recognition have been introduced.Then we analyze and discuss the issues and challenges faced by face recognition technology in the present stage,and the chapter arrangement of the thesis is shown.In the second chapter,the research of face direction recognition based on wavelet neural network is discussed mainly.Wavelet neural network has the theoretical basis of wavelet analysis and it is used in image processing widely.Then we use it for identification and a simple innovation in the hidden layer of the network has been improved.We use other different methods for comparative analysis,and the final results also prove the advantages of wavelet neural network.In the third chapter,a algorithm for face recognition based on ensemble learning is proposed.Due to the incompleteness and inaccuracy in a single face feature extraction and recognition,we propose a face recognition method based on ensemble learning by combining multiple feature extraction and classification integration technologies.The idea of voting integration is adopted at the decision-making level.Then,a fuzzy wavelet network based on feature fusion and LM algorithm is applied to facial emotion recognition in the fourth chapter.In the stage of feature extraction,the time and frequency domain features of face image are fused to obtain more perfect feature information.In the stage of recognition,the wavelet neural network is embedded into the fuzzy neural network to construct the fuzzy wavelet network for classification.The recognition efficiency and accuracy of the model are improved by optimizing the classifier.At the end of this thesis,all the content are summarized and we make prospects for the future with the limited knowledge of the author.