Computers are beginning to be able to process real world objects without the need for codes or tags in their own language. Innovations such as Peres have previously used chemical data analysis to quickly recognize food types and determine if they're ok to eat. Now the Netherlands' Studio diip has created the Vegetable Recognizer, a system that uses image recognition to detect types of vegetables and provide recipes that include those foods.
The idea uses a projection screen that can be overlaid onto any surface to turn it into a smart one. Using a camera located overhead, the system can identify objects that are placed onto it using complex image recognition algorithms. If a zucchini is placed onto the table, the system will confirm that it's detected a new object, identify it, and project its name next to it. Created as a demonstration of the possibilities of image recognition, the platform enables users to place any number of different vegetables onto the surface. The system then checks a list of pre-programmed recipes that contain the ingredients on the table and filters the ones that feature the vegetables available.
Watch the video below for a demo of the Vegetable Recognizer:
As well as helping home chefs decide what to cook with the ingredients they have, the system could also be used in supermarkets to drive purchases of particular items. Are there other ways image recognition could help computers process the real world?