New Lip Reading Technology To Have Both Forensic And Therapeutic Applications

Researchers from the Computer Science department at the University of East Anglia (UEA) have engineered a visual speech recognition system which would act as a support facility for forensic science and help people with visual and speech impairments. Dr Helen L. Bear and Prof Richard Harvey of UEA's School of the Computing Science division have envisioned a technology that will interpret the words by lip reading and can be used whenever the audio is deficit.

Professor Bear explained that manual lip reading is possible but only up to a certain limit. Some sounds such as '/p/,' '/b/,' and '/m/' are significantly similar and might produce an error when interpreted manually. Their team has updated the previous lip reading system and now it is based upon a novel classification system which can aptly differentiate tones and corresponding letters.


The system is still is in its infancy and requires greater study before commercialization. Once prepared, it can be used to filter important words where noise is present or where the sound is not clearly coming from the source. Professor Harvey explained that lip reading has emerged as one of the most challenging problems from the Artificial Intelligence sector and improvising a particular method related to it might address many other tricky problems related to audio-speech recognition.

The research was an integrated part of a 3 year project and funded by the Engineering and Physical Sciences Research Council (EPSRC). The complete research report, titled as “Decoding visemes: Improving machine lip-reading” was published earlier today at the Proceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing 2016.

Source: #-Link-Snipped-#