Research And Implementation Of Face Recognition Technology

Face recognition technology is a research hotspot in the fields of pattern recognition and artificial intelligence.This technology uses machines or computing devices to complete identity verification operations.Traditional face recognition technology is a combination of hand-designed features and machine learning technology,while hand-designed features are difficult to cope with different changes in real scenes.In recent years,face recognition methods based on convolutional neural networks have been widely used by researchers,because deep learning methods can independently learn face representation information through training a large number of data sets,so that they can be applied to changing environments.Face image.At present,the convolutional neural network model trained on the open source data set has been applied to face recognition with high efficiency and accuracy.Therefore,the research on the face recognition technology based on the convolutional neural network has high practical value.This article is mainly to understand and learn the relevant technical knowledge of convolutional neural networks,to study face recognition technology and realize the face recognition framework.First,introduced the key steps of the face recognition system and the technical knowledge involved in each step.Then introduced the theoretical knowledge of convolutional neural networks,exemplified and compared the currently widely used face detection and face feature point positioning algorithms,and passed the research on the network models involved in these two important steps and the network models.Based on the comparison between the results,the algorithm model used in this article is selected,and the network model and source code are studied to optimize the selected network model to realize the face recognition system.Finally,based on the image processing technology,the application of C++11 and the algorithm model proposed in this article,the entire process of face recognition is encapsulated,and a visual interface tool is developed in the Visual Studio2015 environment using the MFC application framework to realize the whole person The flow of the face recognition operation.At the same time,the system framework performance and other aspects are tested to verify the robustness and stability of the tool.