CalMeter: Food Recognition and Calorie Estimation Using Deep Learning
摘要
Humans have three essential needs: food, shelter, and clothing. Of these, food is the most important. Although humans find it fascinating to eat food in a variety of stages, we should only consume what our bodies require. Many serious illnesses have recently been discovered in the human body. The ease with which food is brought to our homes is causing obesity, a serious chronic illness, to rise in popularity. People’s anxiety about their diets increased at the same time that their need for food increased. This is where the” CalMeter” food recognition and calorie measurement method comes in handy. It allows us to upload an image of a food item and identifies the food item and also calculates how many calories it contains. Using deep learning techniques, this model will offer a unique way to measure calories. We assess different deep learning techniques on the novel dataset that contains different classes to see which model will perform optimally. We have trained our dataset on CNN model, VGG16, Inception model and Resnet in which VGG16 algorithm got the best accuracy. With the help of our carefully selected dataset, which includes a wide variety of food photos with accurate calorie labels, CalMeter is able to provide calorie estimations for a wide range of dishes, encouraging people to adopt healthy eating habits.