Use of non-invasive biometric techniques and machine learning algorithms to classify beer quality based on foamability parameters

1University of Melbourne, School of Agriculture and Food; Faculty of Veterinary and Agricultural Sciences, VIC 3010, Australia *Corresponding author: sfuentes@unimelb.edu.au Introduction In recent years, consumers preference and trends for beer have been changing, and are currently looking for more premium beverages. Beer foam and bubble characteristics and dynamics have been identified as the most important quality traits as they are the main visual attributes that contribute to beer quality assessment. Sensory evaluation has been widely used to assess consumer acceptance of food and beverages, however, this method is only capable of obtaining the conscious responses from participants. Biometric techniques such as eye tracking, infrared thermography, facial expressions and heart rate from image analysis assessed through signal and computer vision algorithms, can results in more information related to the unconscious responses. Biometric studies coupled with sensory forms, are able to provide more information about consumer behavior from the conscious and unconscious responses.