Automatic food recognition system for diabetic patients

There is good evidence that eating a healthy diet can reduce your risk of obesity and illnesses such as diabetes, heart disease, stroke, osteoporosis and some types of cancer. The food you eat contains different types of nutrients, which are all required for the many vital processes in our body. Our approach is to capture food and fed to Dense SIFT method, this method extract keypoint and visual vector from an image. Extracted visual vector are clustered using K-means clustering technique. Finally support vector machine classifier is used in this work, classifies the food image and measures the carbohydrate level from food image. Our proposed system is based on Bag of Feature (BoF) model.