Engineers taught Siri and Google Assistant to understand a person without words [VIDEO]

Researchers from Cornell University in the US have developed an original gadget for users of voice assistants Siri, Alexa and Google Assistant. The compact device allows you to interact with virtual assistants without making a sound.

.

A miniature device with the working name SpeeChin is equipped with an infrared camera and is designed to be worn around the neck. The upward-facing lens tracks the position of a person's chin and, after a short period of machine learning, recognizes voice commands, even if the user says them silently. At the same time, the device does not “see” the faces of the owner himself and those around him.

The first test of the technology was carried out on 20 volunteers. They spoke simple phrases, including numbers, in English and Chinese. In the first case, the recognition accuracy was 90.5%, in the second - 91.6%. The indicators worsened if people simultaneously made some movements, for example, walked, because of which their facial expressions changed. The developers claim that this problem can be solved by arranging longer training sessions for the gadget. 

In the future, engineers intend to improve the hardware of the device, which will allow it to more effectively track the movements of the chin by increasing the shooting speed and increased camera resolution. The introduction of this technology in the future may simplify the interaction of users with voice assistants in situations where it is inconvenient to say a command out loud (for example, in public transport).

Share This:

Facebook Twitter Pinterest Linkedin Whatsapp Whatsapp

About Steven Perry

This is a short description in the author block about the author. You edit it by entering text in the "Biographical Info" field in the user admin panel.