Feasibility of food recognition and calorie estimation of fast food and healthy meals available in the Philippines

• Main Authors: Dy, K., Ligan, J., Cabatuan, M.
• Format: text
• Published: Animo Repository 2018
• Subjects: Convenience foods—Caloric content; Food—Caloric content; Computer vision; Neural networks (Computer science); Dietetics and Clinical Nutrition; Electrical and Computer Engineering; Electrical and Electronics
• Online Access: https://animorepository.dlsu.edu.ph/faculty_research/1895
• Institution: De La Salle University

This paper presents the design and development of a food recognition smartphone application which can also display the estimated calorie/s of the food itself. It is intended for people who would like to monitor their diet through food calorie intake measurement (i.e. user's daily calorie intake record). It is equipped with a food database consisting of typical fruits and vegetables commonly found in the Philippines. As part of the study, it also includes some of the meals in food chains (i.e. McDonald's, and The Healthy Corner) found in the Philippines where the calorie information is readily available. The result shows 82.86 % accuracy for the top-1 category, and 99.29 % for the top-5 category. The algorithm being used in this project is Artificial Neural Network (ANN) wherein the recognition process must properly be achieved. Furthermore, the aforementioned database is supported by TensorFlow which is an open-source software library for Machine Intelligence. © 2018 Universiti Teknikal Malaysia Melaka. All rights rerserved.