The Artificial Intelligence along with recognition of food components performs the recognition of the whole dish.
The Whole Dish Recognition page consists of three parts:
For a given dish image the application can propose several names of the dish. For example, the dish with steak, lime and basil can look like the steak and like the tuna salad too.
The app displays several buttons with different dish names. The user can tap one of them, and the button's image changes its color. In our example this is button Steak.
If none of the names in this column satisfies user, he can tap the last button Other... and enter the correct dish name either typing it or pronouncing it.
When user taps on the button with food name, the app refreshes the list of possible components for selected dish.
User can tap several of these buttons to reduce the number of recipes that will be provided to him during database search of the selected dish name. In our example user has highlighted buttons basil and lime. When user taps button Next , the app will perform the search of recipes in the cloud database for dish name Steak and components basil and lime.
If user did not highlight any of components name, the results of database serach will include all the recipes with name Steak
Floating buttons at the bottom of the page:
Tapping button Next user initiates the search of the database for the recipes with selected name. When search is complete the app will open the Selection of Recipe page
Tapping button Speak user gives voice commands to the app.
Button Information opens popup window with brief description of the page, its voice commands and buttons.