Research On The Attraction Of Museum Exhibits Based On Facial Expression Recognition

With the development of the times,more and more people choose to enter the museum,or enjoy the cultural influence brought by the history museum,or feel the intellectual charm brought by the science and technology museum.Therefore,how the museum can better serve the visitors and use the limited exhibition space to meet the visitors’ demand as much as possible has become a very meaningful research direction.Traditional research methods may be mostly limited to questionnaire surveys and human inquiries,making it difficult for visitors and museum staff.Therefore,this article mainly studies how to combine the existing artificial intelligence technology to obtain the visitors’ evaluation of the museum exhibits,so as to facilitate the museum to optimize the display of its exhibits.A large number of studies have shown that some human inner thoughts or emotional states will naturally flow out in his face.By observing and comparing the facial muscle changes brought about by facial features such as the corners of the mouth,eyes,and face,one can perceive their inner emotional state.Therefore,facial expression is an important indicator that intuitively reflects the satisfaction of visitors with museum exhibits.This article combines the prevailing convolutional neural network to realize automatic facial expression recognition,and obtains the visitors’ inner evaluation of the exhibits in a natural way.Combining the traditional eight kinds of expression classification methods and taking into account the actual visitor scenes,this article customizes a set of expression classifications to subdivide the expressions of visitors into three positive expressions: happy,interested,focused and three negative Expression:Confused,tired,bored.In response to this expression classification method,this paper also self-made the expression data set MVE suitable for museum visitors.According to the task of detecting multiple expressions at the same time,the mini＿Xception network was improved to speed up the recognition speed of the model and improve the recognition accuracy.In addition,this paper also combines the facial recognition algorithm based on convolutional neural network to calculate the visitor’s stay time in front of the booth,and uses it together with the expression score calculated by the facial expression recognition algorithm as an evaluation indicator to calculate the value of each exhibition booth in the museum.The attraction of exhibits to visitors.Taking into account the task of distinguishing multiple people at the same time in the actual museum scene,this paper modifies the Face Net network loss function to increase the distance between classes to better distinguish different visitors.Then this article uses the Py My SQL dependency package to import the evaluation indicators into the My SQL database,and then uses the front-end code and back-end joint debugging,and uses the ECharts library in Java Script to realize the visual presentation of the data.Through the visual presentation of these data,it is intuitively reflected whether each exhibition point is enough to attract visitors.Some exhibition points with lower scores can improve the array of museums by replacing exhibits.