Consumer Acceptances Through Facial Expressions of Encapsulated Flavors Based on a Nanotechnology Approach

This paper presents a new methodology for analyzing consumer preferences and acceptance of food flavors through facial emotion recognition. In this study, we applied a method based on nanotechnology to produce encapsulations of several flavor profiles. Facial expressions were detected through the Microsoft Kinect sensor and video images of 120 volunteers tasting five different flavor samples were obtained. A neural network was trained to measure emotions through facial expressions in every frame. Then, the combination of the consumer's evaluations, the frame number interval where the consumers tried the sample, and the expressions found in the videos were used to solve a regression problem using different supervised learning techniques: Support Vector Machines for regression and Multilayer Perceptron and Regression Trees to predict whether a specific taste might be accepted or rejected. We show that this methodology could be used in food marketing.