Design and Development of Computer Vision System for Calorie and Nutrition Measurement

Abstract

Obesity in children and adults has almost doubled since 1980 due to their association with various diseases, such as heart disease, type II diabetes and various types of cancer. Obesity and being overweight is considered to be fifth primary risks for the worldwide deaths. The main reason of obesity is consumption of high calorie foods and absence of physical activities. In 2008, one-tenth of the world’s population was obese and figure is increased to one-sixth in 2012. As a result it is impetrating that we should fight collectively hand in hand against this huge problem before it poses a threat to our health system. Thus, treatments of obesity have been the centre of a huge number of current studies. Because of these studies, researchers have observed that obesity treatment needs consistently monitoring of the patient’s daily diet. However, measuring daily intake of fruits and vegetables is considered a significant step in the success of a good diet. Measuring daily intake of fruits and vegetables for obese patients is one of the most challenging tasks in obesity control studies. Numerous studies have recommended that using modern technology like smart phones may improve the under-reporting problem in daily diet intake consumption. In this thesis, we present a system that helps in calculating the amount of nutrients; we get from fruits and vegetables in daily diet. Our system uses image processing and classification methods to recognize the food and measures the calories and intimates the user about daily/weekly/monthly deficiencies. The evaluation and computation of the fruits and vegetables volume and total calories in the picture is an important step in our system. Our experiment results shows high reliability and accuracy of this method, with less than or equal to 25% error.