The world’s first translation machine, designed specifically for translating British Sign Language (BSL) into written English, is being designed by the Centre for Vision Speech and Signal Processing (CVSSP) at the University of Surrey.
The project, worth around £1 million, plans to develop a machine that will recognise hand motion, form, facial expressions and body language of the person signing. It will then process this information into possible phrases, which are then translated into either written, or spoken English.
The linguistic study of sign language is relatively new, having only started in the 1960s, but it is very interesting to linguists due to what it reveals about the possibilities of languages that don’t rely on sound.
According to CVSSP, the main issue, when studying non-verbal languages for translation, is that it takes a considerable amount of man-hours to complete analysis of video footage necessary for a complete translation. The CVSSP are working towards using computer vision tools to assist with video analysis, which will in turn help with the creation of the world’s first machine readable dataset of a sign language.
The project requires a collaboration of experts from a variety of fields such as, leading experts in British Sign Language, and engineers who specialise in computer vision and machine learning. Together, they plan to use technology such as, data gloves and motion capture systems like Kinect, to create the required data necessary for an accurate translation.
Richard Bowden, Professor of Computer Vision at the University of Surrey, said: “We believe that this project will be seen as an important landmark for deaf-hearing communications – allowing the deaf community to fully participate in the digital revolution that we are all currently enjoying. We are passionate about sign language at CVSSP, so much so that everyone who works in this area within our lab is asked to learn how to sign.”
Professor Adrian Hilton, Director of CVSSP, said “This project builds on a track-record of internationally leading research on visual understanding and translation of sign-language in CVSSP led by Prof. Bowden and his team. ExTOL is an exciting opportunity to realise the first end-to-end system for translation of British Sign Language leveraging recent advances in Machine Perception and AI.”
Professor Bowden’s team recently published a paper describing the AI and deep learning systems that will complete the translation directly from BSL to a verbal language. The paper was presented at the IEEE Computer Vision and Pattern Recognition in Salt Lake City, the top conference for computer vision.
The vast steps, which are currently being made in computer-aided translation, will go a long way to ensuring an inclusive environment for members of the deaf community, and should facilitate a greater integration with digital technology.
Recent Comments