December 2014 Issue
Topics

The FoodCam: Smart phone application for real time food recognition for personalized healthcare

The FoodCam
Keiji Yanai

Keiji Yanai, Associate Professor, Graduate School of Information Engineering, University of Electro-Communications, Tokyo.

The ubiquitous smart phone is becoming more than just a device for communicating and internet browsing. Increasingly, innovators are developing 'apps' for health and medical care such as monitoring blood pressure.

Here, in a significant extension of the applications of smart phones for health care, Keiji Yanai, an associate professor at the Graduate School of Information Engineering of the University of Electro-Communications, Tokyo, has developed the "FoodCam"--a real time, a unique mobile food recognition system.

"By taking a smart phone photograph of food such as a bowl of noodles, the FoodCam enables the user to estimate the number of calories and nutritional value of 256 different types of food," says Yanai. "An internet connection is not necessary. The computation to recognize and analyze the food is carried out by the mobile device itself."

Recognition accuracy is important as stated in a recent publication by Yanai: "Experiments show a 79.2 % classification rate for food from the top 5 category candidates for a 100-category food dataset with the ground-truth bounding boxes when we used HOG and color patches with the Fisher Vector coding as image features".

Plans for future studies include automatic recognition to initiate detection procedures without any user operation, measuring the volume of food, use of 'deep learning' to minimize the memory usage of smart phones, and improvement of the accuracy by using more than a single image.

Publication
  1. Yoshiyuki Kawano and Keiji Yanai, Multimedia Tools and Applications, April 2014
    DOI 10.1007/s11042-014-2000-8
Further information
  1. Yanai Lab. website: http://mm.cs.uec.ac.jp/e/
  2. FoodCam project page: An Android application of the mobile food recognition system and 100/256-class food dataset "UEC-FOOD100/256" can be downloaded from http://foodcam.mobi/
The FoodCam
A screen-shot of the 'FoodCam' and smartphone.