You’re on a diet, and you’re watching calories like a hawk. You check every label at the supermarket and stay away from fast food, but what do you do at restaurants that don’t offer detailed nutritional information on their menus?
It seems like your only two options are to put your diet in jeopardy and take a leap of faith, or avoid the restaurant and miss out on a delicious meal.
A third option is in the works. In the near future, you will be able to take a picture of the food with your smartphone and consult your virtual nutritionist.
Nutritional information web service Asken is using the latest Sony technology to give dieters a handy tool for managing caloric intake. The technology analyzes images of full meals and picks apart the various components, displaying the name and caloric value of each part as well as the total nutritional value of the entire meal.
To be more specific, the technology gathers information on the color and shape of the food and detects how much space it occupies on the plate, then uses artificial intelligence equipped with vast culinary knowledge to estimate what kind of food it is.
That data is fed into Asken’s database, which calculates calories and other nutritional information. Asken pairs that information with advice from nutritionists to help users with their diets.
Take the meal in the picture above as an example. You would take a picture of it on your smartphone, send the image to Asken, and in a few seconds, Asken would return the image with labels of “rice,” “miso soup,” “salt-grilled salmon” and “boiled spinach” and nutritional information for each.
Sony is still developing this technology and has allowed WIT, the company that runs and manages Asken, to run demonstration tests since March 21. There is still a ways to go until this technology can serve as a comprehensive dieting aid; for example the technology is not able to discriminate between two bowls of miso soup with differing caloric content.
“We use the average caloric content of miso soup to arrive at the results,” a Sony spokesperson said. “However, we will collect many images of miso soup in our demonstration tests, so we may be able to use those to make determinations in the future.”
In addition, the technology can only identify a limited number of dishes, starting with foods that show up most commonly.
“[The technology] can identify about 20 foods like rice, miso soup, curry and milk,” the Sony spokesperson said. “We will add more specific items as we continue development.”
Even in its infant stage, this technology sounds like something many people would want to try. Most people who use Asken are dieters, giving WIT an ample supply of participants from which to choose.
“We are unveiling the technology on a limited basis amongst Asken users in this initial stage,” a WIT s said. “We plan to increase demonstration tests as we study the accuracy and effects of the technology.”
Asken is merely the first web service in Japan to use this technology; surely others will follow and attract a multitude of users once it has been improved.
“Our demonstration tests are focused on females ages 25 to 45 because that demographic represents 80% of Asken users,” the WIT spokesperson said. “However, we expect a robust number of male users because this technology makes it easy to keep a record of one’s diet.”