In this seminar, we welcome a visiting academic from the University of Melbourne in Australia, whose work on consumer preferences will be of interest to those working with experimental / eye tracking
approaches to consumer behaviour, albeit his work has had an agricultural / food focus. There will be plenty of time for discussion / Q&A.
“Non-invasive biometric applications for Sensory, Consumer sciences and potential applications in Health, Psychology and Psychiatry”
A novel App (Bio-Sensory App) has been developed by the sensory science group from The University of Melbourne to gather self-reported sensory data and non-invasive emotional and physiological responses
from panellists. The Bio-Sensory App can be uploaded to tablet PCs and different type of sensory scales can be programmed such as 9-point hedonic scale and 3 and 5-points just about right scales, yes/no
questions such as purchase intention, non-parametric continuous scales for descriptive tests, continuous face-scale, and multiple-choice questions for tests such as CATA. Furthermore, the App allows upload
of images, sound and videos as part of questionnaires, which are useful for sensory analysis of design concepts, packaging or pouring of brewages, such as beer and sparkling wine. The App can gather
information for each question from the integrated high-resolution camera from the tablet PC and a FLIR AX8 infrared thermal camera attached to the system, if available. The recorded videos can be processed
using a commercial software, such as FaceReader™ to obtain eight different emotions and two dimensions. With the same videos, heart rate and blood pressure can be obtained using customized machine
learning algorithms developed by our sensory group based on changes in luminosity of different face regions with an accuracy of 85%. Body temperature changes can be obtained using the FLIR infrared thermal camera.
The App is able to upload to the cloud both the conscious responses and videos from participants. The associated analysis system make possible to obtain biometrics from panellists that can be analyzed using multivariate
data analysis techniques and machine learning algorithms to generate models describing liking of food and brewage products based on biometrics / physiological inputs from the App. The system described has been already
tested in sensory booths and in social contexts to assess packaging and different food products such as wine, beer, chocolates and agricultural / horticultural produces. The App has been also tested internationally for trials on
beer and pork. Recently, potential applications have been explored to assess biomarkers and bio-traits for the assessment of Post-Traumatic Stress Disorders (PTSD).
The use of non-invasive biometrics aids in the elimination of bias due to consumers awareness of attached sensors, which may negatively affect the physiological responses from consumers. The App is also able to be integrated
with the use of eye tracking to assess packaging, and images or videos from food products to assess consumers’ acceptability when evaluating a product by their appearance. An example of this may be found in the study “Robotics
and computer vision techniques combined with non-invasive consumer biometrics to assess quality traits from beer foamability using machine learning: A potential for artificial intelligence applications” which concluded that beer
quality is mainly determined by its visual assessment focused on foamability. A study using eye tracking to assess food labels showed that there is no significant difference in consumers’ assessment of visual and physical labels.
Other studies conducted by the group have found strong correlations between the emotional/physiological biometric responses and consumers acceptability of food products such as beer and chocolate. Therefore, machine
learning models have been developed by using the emotional/physiological biometric responses as inputs to predict the level of liking of beer with high accuracy (~84%).
Biography
Dr Fuentes received his undergraduate degree from The University of Talca in Chile in 1999 and his PhD from the University of Western Sydney in 2006. He did two Postdoctoral Fellowships at The University of Technology
in Sydney (Climate Change, elevated CO2 experiments) and The University of Adelaide (Climate Change, Viticulture) then was appointed as Lecturer in Viticulture at The University of Adelaide from 2008-2012. In 2012, he
was appointed as Lecturer in Wine Science and Co-ordinator of the Master of Wine Technology and Viticulture at The University of Melbourne. His main research and teaching interest is the use of state of the art of
instrumentation for plant physiology research, such as short range, airborne and satellite remote sensing, Near Infrared Spectroscopy, Infrared Thermography, Sap Flow Sensors, as well as in the area of computer
programs developed for agricultural research and practical applications, and development of new methodologies to assess plant physiology and growth using image analysis and instrumentation. He is the international
co-ordinator for The Vineyard of The Future Initiative, a multinational collaboration to establish a fully instrumented vineyard for climate change research.
