Public Funding for Liopa Ltd

The COVID pandemic will result in large numbers of patients who need to be weaned off ventilators over several weeks. In order to do this, patients will need a tracheostomy. From experience with existing patients it is known that these patients will not be sedated, will be in an unfamiliar environment, confused, frightened, weak and unable to vocalise. By utilising AI technology the SRAVI app can lip read and then translate to written or verbalised language on a handheld device - allowing these patients to communicate effectively with carers and relatives. The App, when pointed towards the patient, will capture video of the patient mouthing phrases, decipher 'speech' from the lip movements and vocalise the result. This can be done both at the patient bedside or remotely if they continue to be infectious. This will help mental health, pain relief and early rehabilitation. The project will deliver - * A fully-featured, productised version of SRAVI capable of commercial use in ICUs/General Wards. * A Feasibility study using the SRAVI App in two (or more) ICUs - Usability measures will be built into the app to capture feedback and allow ongoing development and evaluation of the app. A range of outcome measures will be collected from patients, families and staff to inform the healthcare impact and market potential of SRAVI. In a follow-on phase the productised version of SRAVI that will be completed at the end of the on-site Feasibility Study will be prepared for commercially release - and then commercially launched on the Android and Apple App stores. This commercial version will then be evaluated in a clinical setting at Royal Preston Hospital. In addition the Liopa team will implement internal systems and processes for the future development and operation of the SRAVI service that conform to the regulations required to obtain Class 1 Medical Device certification.

"There are a group of people who are unable to vocalise or move their limbs but are able to move their lips. In a hospital setting these include those with neck and spinal cord injuries, degenerative neurological conditions and some head and neck cancers. There are only a few ways of allowing these patients to communicate and those resources are expensive (lip readers or 'eye gaze' devices) and need extensive training to use. Liopa are proposing the **SRAVI** application (**S**peech **R**ecognition **A**pp for the **V**oice **I**mpaired) which will be based on **LipRead**, Liopa's Visual Speech Recognition platform. This novel application of a proven technology allows translation of lip movement to text using an android app on a mobile device which will need very little training and is inexpensive. It uses artificial intelligence to learn how an individual's lips move and stores this function in the cloud. This allows various devices to access the system unlike ""eye gaze"" systems which are restricted to one device per affected individual. This also allows communication between patients, staff and family from an early stage. This empowers the patient to take a more proactive role in their care and recovery improving wellbeing and potentially reducing the length of recovery in certain cases. This feasibility study will look at a select group of patients with tracheostomies (approximately 10 000 tracheostomies are placed per year in the UK) who currently struggle to vocalise but can move their lips to establish whether the software is accurate and if it is usable in the clinical setting of a critical care area of an NHS hospital. Starting from a limited vocabulary in English, our software has the potential to be expanded to become much larger and apt to use in other languages and markets."

To develop a Visual Speech Recognition system, capable of recognising a limited vocabulary, which can act as an enhancement to an audio speech recognition application such as in car voice control.