Food calorie estimation using machine learning and image processing

In today’s world a healthy lifestyle is a must for every individual and what they consume is of utmost importance in order to achieve the same. Our paper focuses on creating software which gives the calorie of the food which the user is going to consume. In order to achieve this, the software will take two images as input from the user, the top view and the side view. The image will have a probe object as well which is a coin whose volume will be known. The food item in the image will be detected with the help of Faster R-CNN algorithm. We are using Faster R-CNN algorithm since it is the fastest among all the object detection algorithms. In Faster R-CNN we don’t use selective search algorithm on the feature map to identify the region proposals, we use a separate network is used to predict the region proposals. This makes it the fastest algorithm and therefore it can be used for real-time object detection. In the next step, we will segment the image using grabcut algorithm. It is needed for foreground extraction with minimal user interaction. After segmentation of images, the volume of the food item is calculated using the known volume of the probe object. After the calculation of volume, the mass of the food item is calculated with the help of formulas and then the calories of the food item will be calculated using the relation between mass and calories.