Individual recognition of food in bulk by using 3D model of food

In this paper, we propose a method of individual recognition of food in bulk by using 3D model of food.First, color images and depth images of them are generated by using 3D model of food and physics engine of simulator.Then, color and depth composite images are created by converting two channels from color images and one channel from depth images.In the experiments, the accuracy of individual recognition of food in bulk with color and depth composite images are shown to compare the accuracy with only color images.