Anantha Krishnan Raman Fesabas, David Asirvatham & Jean-Pierre Poulain
Abstract: The integration of technology in various sectors, including food studies, continues to increase exponentially. Computer vision, a research topic which initially started with object detection, facial detection, followed by facial recognition (FR), is currently focused on the more complex facial expression recognition (FER). The present research explores FER and its relevance in food studies. In addition to examining some of the current FER methods, such as Convolutional Neutral Network and Haar Cascade, the research is also developing a pre-processing method to increase the accuracy of FER results. FER methods are generally designed to study customer feedback and reactions towards meals served in the F&B industry. Typically, cameras are set up at different corners of a restaurant to capture facial expressions for analyses. Although one can never truly identify a person’s feelings just by expression recognition alone, together with survey and feedback forms, researchers are able to get more reliable and accurate data with regard to a customer’s preferences and taste.
Keywords: Computer vision, facial expression recognition, visual image noise optimisation, VINO, food studies