June 12, 2017
A group of Brown University students has developed a smart kiosk called Zest that uses computer vision to recommend spice combinations for cooking.
The user places their food in front of the kiosk, and a screen shows spice recipes for the dish. As the user explores the recipes, the recommended spices appear on an adjacent spice rack.
The kiosk was developed for the students' engineering class.
The students learned through research that it is hard for most people to know what spices to use in recipes. They developed a system to provide a limited number of recipes, saving the user from having to search for recipes and buying the spices.
The system has a camera and a 7-inch LCD touchscreen to detect specific foods and present spice recipes. When the food is put in front of the camera, it sends an image to a Google web app which the students built. The web app sends the image of the food to a pair of third-party image recognition services that return a set of strings describing the foods in the image. The food names are queried against a database of spice recipes.
Designed for college dining halls, Zest helps students with limited experience using spices learn about combining spices.
The kiosk can be integrated into university dining halls or any space for communal dining.
The process of developing the kiosk is explained in detail on the Zest website.