Scientists from University of East Anglia (UEA) have created multi lingual lip-reading computers that can recognize what language you speak by tracking face and lip movements. It will be useful for deaf people, law enforcement agencies and those who are in a noisy environment. Though computers that can read lips are nothing new, this is the first time computers have been taught to recognize different languages.

At the University’s School of Computing Sciences, researchers developed the ground braking technology by statistically modeling the lip motions of 23 bilingual and trilingual speakers. So if an individual speaks any language, the lip reading computers can identify the language with “very high accuracy.”

It can identify the following languages: English, French, German, Arabic, Mandarin, Cantonese, Italian, Polish, and Russian.

So how does it recognize different languages? When you speak different languages, you use different mouth shapes in different sequences. For example, computer vision will recognize the French language from “lip rounding” movement and Arabic language from “tongue movement”.