Public Funding for Signapse Ltd

Signapse wants to make as much information as possible available in sign language. In this project, we are building AI software that automatically translates from written English text to British Sign Language video, delivered by a photorealistic digital signer that is indistinguishable from a human. This will ensure the Deaf community can have the same access to information as their hearing peers. British Sign Language (BSL) is used by more than 150,000 people in the UK and is the preferred language for at least 87,000 Deaf people. As most Deaf people grow up with English as their second or third language, ordinary information, whether found online or in person, is mostly inaccessible to them. This seriously impacts their access to healthcare, employment, education, transport and more. The current solution to this is to rely on BSL interpreters. However, with approximately 1 interpreter to every 100 Deaf individuals, as well as barriers like cost and geography, it is evident that the needs of the Deaf community cannot be fully met. Our technology can complement the interpretation profession, making it both straightforward and cost-effective to deliver sign language translation automatically and at scale. Having successfully developed AI translation for constrained information settings such as train and flight announcements, our project focuses on working with translators to build a large-scale BSL dataset, and using this to train an AI model for automatic text-to-sign translation. Signapse was co-founded by Sally Chalk, the founder and ex-CEO of Clarion, the UK's largest BSL interpreting agency; Professor Richard Bowden, Professor of Computer Vision and Machine Learning and Head of the Cognitive Vision Group at the University of Surrey; and Dr Ben Saunders, who holds a PhD in Sign Language Translation. Working with our Deaf Leadership Board and in consultation with the Deaf community, we ensure that the needs and preferences of Deaf people are understood and consistently reflected in the development of our technology.

Signapse uses AI technology to bridge the communication gap between Deaf and hearing communities. We are building software that automatically translates from written text to Sign Language video delivered by a photo-realistic digital signer, indistinguishable from a human. Our mission is to ensure the global Deaf community can have the same access to information as their hearing peers. There are an estimated 70m Deaf citizens globally and 300 sign languages in use around the world. British Sign Language (BSL) is used by more than 150,000 people in the UK and is the preferred language for over 87,000 Deaf individuals. For most of them, English will be a non-native second or third language. By using our AI technology to dramatically increase the amount of information available in sign language, we can improve the access that Deaf people have to key services, like transport, in a way that represents significant value for money. We want rail travel to be accessible to Deaf passengers at every stage of their journey. The goal of our project is to create a mobile-based solution which allows Deaf passengers to access national rail timetabling information, in-station announcements and on-train journey information in their native language: British Sign Language. By making this available through API, our solution can be integrated into any digital platform; any train operating company or rail organisation can then utilise this to improve access for their Deaf passengers in a way that is cost-effective and does not require them to procure additional materials such as display screens or subscription to an additional app. During this project we are working with Whoosh, a transport technology innovator, to make our solution available on-demand within their own mobile dashboards; these are accessed via QR codes displayed either in-station or on-train. We are supported by our partners Network Rail and South Western Railway in the testing and demonstration of our first-of-a-kind solution.

There are approximately 87,000 people who use British Sign Language (BSL) daily in the UK. Only 1% of websites are considered accessible to these people (RNID, 2018). Deaf people often have poor literacy rates and struggle to understand written English. The traditional solution is to use a BSL Interpreter to enable Deaf people to understand important information. However, there are only 1000 interpreters in the UK; meaning a lengthy wait for interpretation of about two days on average. The result is a communication barrier between public services and Deaf people that need access to information, with a particular struggle with healthcare. Research has shown that Deaf people's health is already much poorer than hearing people's, with accessibility to information a major factor in this (SignHealth, 2020). Accessibility is even lower online, with almost no examples of signed content on healthcare websites. The main reason is the cost of hiring interpreters to sign all the content, with a £180 an hour translation cost infeasible. In this project, Signapse will scope the feasibility of building automatic sign language translation technology for healthcare websites. Backed by 20 years of computational research, Signapse's mission is to bring AI and sign language together. Signapse was founded after being approached by major organisations that wanted a real-time and cheaper alternative to BSL interpreters. We have developed world-leading innovative technology that is well received by the Deaf community. Our next development phase is to investigate innovative translation methodologies for constrained domains such as healthcare. Healthcare websites are an ideal place to develop our translation technology, as it is both a useful application but also a constrained domain which requires less data to tackle. Within the proposed project, we shall conduct a feasibility study to assess the requirements of the technology, understand the task and enable a commercialisation plan for the market. Due to the large degree of innovation involved with this technology, we believe it is important to first scope the market before we catalyse further innovation on our path to commercial success. If this project is successful, it will enable more accessibility for Deaf people online, meaning they can better understand healthcare content and access self help materials online, particularly pertaining to mental health and wellbeing. This will increase their ability to perform self-care, without relying on human interpretation. We believe this project will significantly improve the health of all BSL users.