A technology that allows you to speak from an avatar just by thinking with your head has appeared, and a woman who can not speak due to a stroke regains her voice and facial expressions with the power of 253 electrodes and AI
A research team at the University of California, San Francisco has developed a technology that allows avatars to speak their thoughts by embedding electrodes in the brain and connecting them to a computer. Avatar can also reproduce facial expressions, and a woman who has lost her ability to speak due to a stroke has succeeded in regaining her voice and facial expressions.
How Artificial Intelligence Gave a Paralyzed Woman Her Voice Back | UC San Francisco
Mr. Ann, who lives in Canada, worked as a teacher, but after suffering a stroke in 2005, he suffered from ' confinement syndrome ', in which he was conscious but could not move his muscles. Mr. Anne has been able to move his body slightly and speak simple sounds through rehabilitation, but it continues to be difficult to communicate without relying on equipment. For this reason, Mr. Ann communicates using a `` device that selects alphabets by eye line and writes sentences '', but conversations using this device take several minutes per sentence.
A research team at the University of California, San Francisco has developed a technology that allows avatars to speak their thoughts by embedding 253 electrodes in the brain and connecting them to a computer. Electrodes are embedded in the brain's ``areas that are activated during conversation'' and ``areas that are activated when changing facial expressions''. You can apply dialogue and facial expressions to your avatar by converting.
The connection between the brain and the computer looks like this.
By using this system, Mr. Ann was able to communicate more quickly than before.
You can see how Mr. Ann communicates through the avatar in the following movie.
How a Brain Implant and AI Gave a Woman with Paralysis Her Voice Back-YouTube
It is said that the developed AI is trained to read ``the composition of the sound you are trying to utter'' rather than ``the word you are trying to utter''. For example, the word 'Hello' is recognized as a combination of four sounds 'HH' 'AH' 'L' 'OW' and then reconstructed as a word. By introducing a system that reads this sound structure, it has become possible to convert words three times faster than systems that try to read words directly.
In this research, the brain and computer are wired, but the research team aims to build a similar system with wireless connection in the future.
Related Posts:
