You can command Siri without making a sound! The Chinese team of Cornell University developed the silent language recognition necklace. Chinese and English are OK

Jin Lei From the Aofei temple
qubits | official account QbitAI

this , Maybe now with Siri communication The most fashionable The way of the ——

It only needs Move your lips , Can make it under Yours control.

you 're right , Namely No need to make any sound That kind of .

This is the latest research result from Cornell Chinese team  SpeeChin, Silent speech recognition .

like this ：

Without sound , Can you guess what he is saying （ The answer is revealed at the end of the article ）？

But in SpeeChin blessings , current Siri、Alexa You can already identify , It also supports Mandarin and English ！

There is a kind of “ Silence is better than sound ” I feel it. .

Shut up , How to control voice assistant ？

Developed by Cornell team SpeeChin, It's actually a magical The necklace .

Put on this necklace , You can open the door to silent voice control .

For example, recite a sentence silently in front of the computer “Next”, The music will automatically switch to the next song ：

Maybe you will say that a word is too simple , Don't worry. , Now let's make it harder ：

“Switch to Taylor Swift.”

so ,SpeeChin Even in the face of a complete script , It can also be accurately identified .

And this one “ Silent recognition ” Key to Necklace , It lies in the... Installed under the necklace infrared camera .

It can capture deformed images of people's neck and facial skin , So as to carry out analysis and identification , Determine from “ Mouth ” What command is conveyed .

And this necklace can not only recognize English silently , even chinese  （ mandarin ） Can also hold live .

In the initial experiment , Yes 20 Participated in the test , among 10 People speak English , in addition 10 People speak Mandarin .

The researchers asked the subjects to give dozens of instructions in English and Chinese , Including digital 、 Interactive instructions 、 Voice assisted instruction 、 Punctuation commands and navigation commands .

The final experimental results ,“ Silent recognition ” The accuracy of the method has reached 90.5% and 91.6%.

The Chinese team from Cornell University

The study comes from Cornell University , One is Ruidong Zhang.

among ,Cheng Zhang Is the corresponding author of the study .

He is an assistant professor in the Department of information science at Cornell University , His research focuses on how to get information about the human body and its surroundings , To meet real-world challenges in various application fields , Such as interaction 、 Health perception and activity recognition .

A work of research Ruidong Zhang, He is Cheng Zhang One of the professor's doctoral students , Graduated from Tsinghua University .

His research interests lie in human-computer interaction and pervasive computing （ubiquitous computing）, In particular, the establishment of new wearable devices , To promote interactive experience and deepen our understanding of human behavior .

As for why they do SpeeChin This work , According to the Ruidong Zhang explain ：

Mainly considering that in some specific occasions , People may not be suitable for making sounds ; Also consider some people who can't speak .

Last , To reveal the beginning GIF The answer ：

Are you right ？

Reference link ：

[1]https://techxplore.com/news/2022-02-smart-necklace-silent-english-mandarin.html
[2]https://www.youtube.com/watch?v=pDf86dSPYYI
[3]https://infosci.cornell.edu/content/zhang-0
[4]https://infosci.cornell.edu/content/zhang-1