Researchers of Intel Corporation released a new software license in the framework of an open-source. It allows the construction of computers that can “Read lips” and see how people developers. In this way, PCs can understand spoken commands.

Current powerful speech recognition algorithms work satisfactorily if no background noise or a very good set headset is used. Their accuracy can be very quickly after when the applications with quite usual high-noise environments need to cope, as is the case in public places. In combination with facial recognition algorithms of Intel’s OpenCV computer vision library allows the use of audio visual speech recognition (AVSR) software, the computer to recognize the face of a spokesperson and follow their mouth movements. The synchronization of video data with language identification makes it possible to significantly more accurate speech recognition and thus improves a wide range of voice applications in high-noise environments. The AVSR software is part of the Intel OpenCV computer vision library. This collection consists of more than 500 image processing functions that enable researchers to develop image-to-text applications.