Artificial Intelligence & Machine Learning

Microsoft’s Smart Fabric Detects Objects Placed On It

The new smart fabric knows what’s on the table.

Imagine you forgot your keys on the table before you left home and your table cloth is reminding you to take them with you through your phone. How cool would that be? 

Researchers from Microsoft have developed a new contact-based object recognition technique called Capacitivo that can turn regular fabrics into interactive ones. Smart fabrics enhanced with Capacitivo can identify the items or foods placed on it thanks to the developed algorithms. The smart fabric also works with glasses, cans, or containers including the liquids inside them. 

How? The smart cloth uses the combination of a capacitive electrode grid along with machine learning to identify the material and shape of a given object.

For testing, the Microsoft team used a prototype of a 12 x 12 grid of electrodes, attached to a conductive fabric. Out of 10 people that have used 20 different objects, the project resulted in a 94.5% accuracy rating so far. 

The algorithm of Capacitivo can suggest a meal to the user after identifying the foods on the table or offer a drink recommendation. The project, according to the researchers, was mainly designed for memory aid in case you forgot your keys or earphones before leaving home along with aiding in the kitchen for cooking.

“This paper demonstrates a technique for contact-based object recognition on interactive fabrics, enabling them to detect everyday objects. We first discussed our sensing principle and investigations into the size and separation between electrodes that were optimal to enable our approach. We then built a prototype with a 12 x 12 grid of electrodes made from conductive fabrics,” said the researchers about the study. 

Thus, researchers decided to incorporate Capacitivo into the table cloth, making it easy to use and practical both at home and/or at the office for all purposes.

Source:https://interestingengineering.com

Leave a Reply

Your email address will not be published. Required fields are marked *