Kindergarten-Based Emotional Analysis System And Key Technology Research

With the improvement of residents' living standards,more and more parents are aware of the importance of child education.However,in recent years,child abuse has occurred in kindergarten teachers.Child education and safety issues have become hot topics in society.At the same time,the development of kindergarten informatization needs to be urgently improved.The current system cannot monitor the safety of children in real time and respond to abnormal events in a timely manner.It also lacks objective data for judging the effectiveness of the curriculum and cannot effectively monitor it.Emotional data of young children can be used not only for enlightenment education,but also forjudging the safety status of young children.The application prospect is broad,and it is an important development direction of intelligent child education.This article proposes to design a kindergarten-oriented emotion analysis system and researches the key technologies involved.The specific work content can be divided into the following two parts:First,based on the analysis of system requirements,four functional modules of the system were designed and implemented:information entry,child analysis,curriculum assessment,and class management.The information entry module is responsible for collecting information for identifying identities;the child analysis module can view the daily mood data of the child in real time,and supports system notification of abnormal emotions;the curriculum assessment and class management module can quantify and analyze the daily course effects and each class management quality.The above modules are coupled with each other to jointly realize the kindergarten emotion analysis system,which helps the gardener to conduct more efficient management and higher quality personalized education.The second is to study the key technologies involved in the system,including "combination of cloud and edge",improvement of face recognition accuracy,and children's emotion data mining.First,the emotion analysis of young children involves the transmission and processing of massive pictures,and the cloud computing paradigm will cause higher system delay.This paper proposes a system architecture combining edge computing and cloud computing,which successfully reduces the system delay.Secondly,due to the lack of children's characteristics and the low pixels of each child in the kindergarten scene,the accuracy of face recognition based on static pictures is insufficient.This paper proposes a facial identity correction model based on location features,and proves the superiority of the model through experiments.Finally,this article analyzes and mines the real emotion data obtained from kindergartens,and explores the relationship between infant emotions and infant safety,infant education.Through the analysis of children and teachers'emotions data,and according to the guidance of experts,it provides a way of curriculum assessment and class management.Based on the above research,this article has developed a kindergarten-oriented sentiment analysis system and put it into use in kindergartens,which meets the needs of monitoring child safety and evaluating child education,and has received good feedback from users,proving the effectiveness of the system.