Seeing-eye spectacles

A new technology promises to be a big help to the visually impaired for a multitude of tasks.

By ADAM GONN
OrCam521 (photo credit: Courtesy)
OrCam521
(photo credit: Courtesy)
Reading a menu or going grocery shopping are simple everyday actions for most; but for the visually impaired, these tasks can be difficult and occasionally even embarrassing, particularly when having to request assistance. Now, however, with the help of a camera, a bone-conduction speaker and a small computer developed by the OrCam startup, these and many other daily actions should become a lot easier for the visually impaired.
The camera is small and light enough to be attached to the user’s regular glasses, or sunglasses, and is connected via a thin cable to the processing unit, which the user can carry around in a pocket or on a belt clip. While the software behind the product is complicated, it is very user friendly when it comes to usage.
All the user needs to do is to point at the object that needs to be identified and the artificial intelligence of the processing unit understands what the user wants to know. For example, if the user points a finger at an item on a menu, the system will understand that he or she wants the text on the menu to be read and promptly proceeds to do so.
Additionally, the range of the camera isn’t limited to objects close to the user; it can also identify if a traffic light is green or red, the name of a street, or the number of an approaching bus. All that is required of the user is to point a finger at the target. And because pointing at random people in the street is not always advisable, some functions, like face recognition, are always running in the background. To add a new face, simply wave a hand in front of the face.
The camera, like the human eye, requires light, so its function is limited when it’s dark outside or under artificial lightning. In addition to recognizing menus, buses and faces, the user can also teach the system to recognize new items, such as packages of food at a supermarket.
There is no need to tell the OrCam to add a new object; all one needs to do is to point a finger at the item and the camera will identify it.
If the camera doesn’t recognize the object, it will say so, and by keeping one’s finger up, the camera will understand that a new object needs to be added. The system will then ask the user to hold the item in front of the camera so that the item can be recorded from several different angles. Thereafter, the system will ask the user to read the name of the object, which will be read back to the user the next time the object is placed in front of the camera.
So, for example, if the user wants to be able to go shopping at a local supermarket, this function becomes very handy. After a number of initial visits during which some time will have to be spent adding new items, the user will then be able to go to the store alone, like any regular customer, without having to ask a store clerk or a friend to come along just to help read the labels.