Estimating Food Calories for Multiple-Dish Food Photos

Abstract

A food photo generally includes several kinds of food dishes. In order to recognize food images including multiple dishes, we need to detect each dish in food images. Meanwhile, in recent years, the accuracy of object detection has improved drastically by the appearance of CNN. In this paper, we apply Faster R-CNN [10], a major object detection method, to food photos of multiple dishes. In the experiments we verify by using two kinds of food photo datasets. In addition, this food detector is applied to food calorie estimation for food photos of multiple dishes. We use Faster R-CNN as a food detector to detect each dish in a food image, and the food calorie of each detected dish are estimated by image-based food calorie estimation [2]. In this way, we estimate food calories from a food photo of multiple dishes. In this experiment, we collect food photos of multiple dishes with total food calorie of multiple dishes. Then we estimate food calories from food photos of multiple dishes by combining the food detector and image based food calorie estimation.