A Real Time Approach to Recognize Facial Expression Based on Scoring System for Restaurants

In this modern era the Automated and unmanned restaurants have started witnessing a surge in demand and popularity, however, customers feedback and any details regarding their experience at the restaurant are hardly able to be registered in such systems. To this issue, this outlines a novel method of scoring customer satisfaction through facial expression recognition that integrates a deep learning model based on Efficient MobileNet for feature extraction and Long Short-Term Memory for emotions time sequence analysis. This hybrid method takes into account both topological and chronological aspects of facial motions. The system consists of web applications for users, a web server for receiving requests and an AI server that receiving, evaluates the image of a person's facial expression and classifies their emotional state into three categories-satisfied, neutral and disappointed so it can be used to understand feedback about food quality and the dining area. Also, improved effectiveness is achieved by using MobileNet and LSTM the fusion is lightweight while being accurate for real time facial expression recognition.