Design And Implementation Of Face Recognition Smart Lock Based On Neural Network

With the continuous development of science and technology,smart locks for face recognition are more and more popular with people and add luster to modern life.As an emerging technology,face recognition has flourished in many biometric fields due to its unique advantages,such as the uniqueness of the face,and it has played great value in many fields.As people’s quality of life and safety awareness increase,the requirements for smart locks are also increasing.More modern and intelligent face recognition smart lock systems can meet people’s needs.This article mainly studies how to make better use of deep learning tools to apply face recognition technology in the engineering of smart locks,especially in the application environment based on Asian faces.The main research contents are as follows:First,the functions and overall structural framework of the face recognition smart lock system are designed.The Tensorflow deep learning learning framework is used as the development platform.Designed and implemented a video stream capture module,face detection module,face pre-processing and face recognition module using Java language as a tool;then,use Face Net deep learning network model to train the face dataset,complete the implementation of the overall system,and transfer learning of Asian face data on the trained face model;finally,the model is quantized and compressed,and a face model suitable for practical engineering is obtained.Experiments show that using the friendly arm of the NanoPi M3 development board and Android as the operating system,face detection and recognition can be well achieved.