Improving the accessibility of content for the Deaf community is an important goal for both EU governments and broadcast industry regulators. Although legislation is being used to coerce content producers and broadcasters to do so, the cost of producing sign-language content and the negative impact of having a sign-interpreter appearing on the content for hearing individuals has relegated sign-language programming to late nights or a small number of sign-presented programs. A low-cost solution for personalized sign-interpreted content creation can address both of these problems, leading to greater accessibility to media content for Deaf users. CONTENT4ALL proposes such a solution to the problem in the short-term, which is also commericalizable, and proposes innovations to technologies that can lead to automated sign-translation capabilities in long-term. To this end, it builds upon the technologies demonstrated by the consortium partners in previous EU projects. As a first development a remote signing apparatus (located off-premises of the broadcaster) will be produced to capture a human sign-interpreter’s signs, pose and facial expressions and to parameterize this information. Then, it will be rendered photo-realistically as a 3D representation of the human sign-interpreter at the broadcaster for production of the personalized stream for Deaf users. While this solution can be used commercially, resulting datasets will provide a vast source of information for learning how to parameterize sign information for translation purposes. The second development of CONTENT4ALL will focus on advancing the algorithms/models used to do so, with the intent to create an open dataset for further research into automated signing. Finally, CONTENT4ALL will demonstrate automated sign-translation applied to a real-world TV broadcasting scenario, which is envisaged to lead to new approaches and innovations in the area in the long-run.

SHERPA will empower interventional radiologists (IR), an overburdened group of specialists who use medical imaging and image-guided devices to perform complex, high-risk interventions. Since trust in technology is rooted in relationships – not in a technical specification or feature, AI-powered assistive technologies will be delivered as a seamless, trusted companion (a ‘sherpa’) across the workflow for two clinical domains (Neurology and Oncology), validated with IR and patients through seven clinical studies, and made available to other medical specialities through a framework methodology and outreach to the entire IR community. The automated workflows and their benefits at user, patient and system levels will be evaluated for two clinical applications: 1. Minimally invasive neurovascular interventions (brain aneurysms) and 2. Minimally invasive interventional oncology (liver tumour ablations). Fueled by cutting-edge advances in AI and robotics yet profusely human-centric, workflow automation will minimise the risk of errors and inconsistencies, ensuring the accuracy and reliability of clinical decisions and boosting IR confidence and job satisfaction. As repetitive, time-consuming tasks can be performed automatically, the IR will gain the much-needed relief to focus on the intervention and the interaction with the patient; the expertise threshold will be decreased, improving workload distribution and team dynamics. Through generalising principles and methodological framework, the insights from SHERPA will extend their benefits to other medical specialties dealing with complex workflows and critical decision-making.