Seeing Eye Shoes Get Upgraded with AI to Help the Blind and Vision Impaired Navigate

Austrian startup Tec Innovation has developed its Innomake shoes for the blind and vision-impaired persons that warn the wearer of obstacles in their immediate path. In their current form, the shoes have been approved as a medical device and are outfitted with toe-mounted ultrasonic sensors that can detect objects up to four meters away. When an object is detected, the shoes will vibrate and sound an auditory alarm, warning the wearer to change their path. While the Innomake shoes are beneficial to those wearing them, they can’t detect what type of object is in the way, or its orientation, critical pieces of information, especially if there are holes, walls or stairs.

To improve the shoe’s obstacle detecting capabilities, the engineers at Tec Innovation reached out to computer scientists at the Graz University of Technology to help develop a camera-based edition to its ultrasonic sensor system. The team at TU Graz has developed deep learning algorithms modeled on neural networks that are capable of recognizing objects and making sure the area is safe to walk through. The algorithms were trained using machine-learning applications and can analyze images taken from the shoe-mounted cameras to identify object types and their locations.

What’s great about incorporating AI into the Innomaker design is that the algorithms can be operated on specially designed mobile systems, as the latest processor technologies can handle the heavy data analysis required for object detection. Tec Innovation is currently working on integrating the system into the shoes, which provides a color-coded overlay to determine the safest path to walk. The engineers at Tec Innovation are looking to take all of the information from the ultrasonic sensors and the AI to create a street-view navigation map for the blind and vision impaired. The company has submitted an application to fund this development to the Austrian Research Promotion Agency FFG.