Computer Vision to Help the Blind

Blind man reading

Image Source: Eyra Ltd
In recent years, cars and drones have been able to make technological advances by combining machine learning and computer vision. However, that same technology can be applied elsewhere in order to dramatically improve the lives of those with visual impairments.

Eyra is a start-up company looking to use machine learning and computer vision to do exactly that with the help of a wearable technology called Horus. The starting point of this technology is a pair of Aftershokz headphones that has a camera module attached to the right side. This module has two Full HD mobile cameras that hang off the side of your head. This headset is connected through a micro USB cable to a black plastic box with a battery inside. The battery has enough power to sustain the device for a full day. Unlike other comparable visual aids, Horus processes information locally so it is not dependent on a wireless connection.

At the moment, Horus is designed to read text, identify objects and recognize faces of the people you are talking to. There are 3 buttons on the black box which will let you select which of the 3 actions is to be performed.

This start-up, which was created by students from the University of Genoa, was originally intended to be a computer vision system for robots to navigate. During their research process, they found that this technology could be applied in different ways, thus leading them to start development of a portable version of their original technology that will provide more aid to those with visual impairments.