Research And Implementation Of Anthropomorphic Customer Service Face Generation System Based On Micro-Expression Optimization

In many public service venues,user service experience is an important indicator of customer service quality.Intelligent service robots can undertake repetitive and standardized business in the business hall,effectively divert customers,and can greatly improve the service efficiency of the business hall.Therefore,how to improve the natural and anthropomorphic experience between customers and intelligent anthropomorphic customer service has become the key to improving the service level of business halls.To meet this demand,this paper proposes a facial micro-expression generation method based on generative adversarial networks.The method realizes overall facial micro-expression generation through a cascaded progressive facial expression generation model,while additionally adjusting the position of the iris in the eye.Using image processing technology to judge the positional relationship between the iris and the eye and calculate the gaze direction of the pupil,and through the adj ustment of the human face action unit,the generation of different facial microexpressions is realized on the basis of the results of the proposed eye adjustment model.The comparative analysis experiments on the data set prove that the model proposed in this paper can effectively adjust the direction of eye gaze and modify facial micro-expressions.At the same time,lip synchronization is very important to improve user experience.In order to meet the user’s need for timeliness and improve the usability of the system in the actual production environment,this topic proposes a rapid face video generation method based on lip synchronization.By using the dynamic time warping algorithm to calculate the Mel cepstrum coefficients of the two audios,the similarity between the audios is obtained,and the generated video clips are dynamically intercepted and merged to realize the rapid generation of intelligent anthropomorphic customer service images.Through experiments on speech and video of different lengths and similarities,the model proposed in this paper can effectively improve the generation speed of lip-sync videos,thereby improving the timeliness and user experience of intelligent anthropomorphic customer service.On the basis of the experimental results of the algorithm model realized in this paper,this paper designs and implements a face generation system for intelligent anthropomorphic customer service.Functions such as business data management and front-end visualization,so as to strengthen the company’s customer service capabilities,and promote the standardization and standardization of consultation in the business hall.This article first elaborates the research background and significance of the topic,and further introduces the research status at home and abroad and the related technologies that need to be used in the research process.Then,according to the application scenario of the intelligent anthropomorphic customer service face generation system,the requirements of the system are analyzed,and the method research and experimental verification are carried out on the key issues.Next,the overall architecture of the system is designed and implemented,and the deployment and testing of the system are completed.Finally,we summarize the work of this topic and prospect the future research.