Face Expression Recognition System Based On Machine Learning

With the development of large data and artificial intelligence,face recognition has made good progress.Various algorithms can recognize faces in images with high recognition rate.With the progress of basic theoretical research,face recognition has been applied in various scenarios.However,with the deepening of application,face recognition has shown some limitations,such as the actual scene in addition to face recognition,but also want to get some more in-depth information of the subject,such as gender,age or facial expression under test,and so on.Among them,facial expression recognition technology has a wide range of scenarios and market space,including human-computer interaction,intelligent control,public safety,intelligent medical and other industries.In this paper,a facial expression recognition system is designed for the application of facial expression recognition technology in intelligent medical care.The system can detect the real-time images and judge whether there is a face.If there is a face in the picture,the facial expression can be classified and the recognition function of facial expression can be realized.The main contents of this paper are as follows:1.Face Detection under Unconstrained ConditionsThis paper mainly introduces the principle and application of normalized pixel difference feature in face detection algorithm.AdaBoost algorithm is used to pick out the most distinguishing feature and form a strong classifier to realize the function of face detection under unconstrained conditions.2.Feature extraction of facial expression recognitionThis paper lists various methods of facial expression feature extraction based on static image and dynamic image,then compares and analyses various algorithms of facial expression feature extraction,and chooses the convolutional neural network algorithm which is more robust and accurate as the feature extraction algorithm of facial expression recognition in this system.3.Research and Implementation of Face Expression Recognition AlgorithmsThis paper introduces the expression database shared by the network,and makes the expression database based on the analysis of the application scenarios of the system.By using convolution neural network model to extract facial expression features,facial expressions are classified and recognized,while guaranteeing low time-consuming and preventing over-fitting.4.Realization,optimization and testing of facial expression recognition systemVarious functions of the detection and monitoring end of the system are realized,and the inefficiency of traditional expression recognition methods is analyzed according to the practical application.The problems of low accuracy,weak robustness and long time-consuming are solved by the convolutional neural network optimization scheme of the system,which can meet the needs of the practical application scenarios of intelligent medicine.At the end of this paper,it summarizes the innovation and research results of the system,and puts forward the prospects for the shortcomings and the areas to be improved.