Innovation Challenge Fund (ICF) d.com



Innovation Challenge Fund (ICF)Concept Note1.Name / Organisation Manish Arora, Indian Institute of Science, Bangalore2.Email address / Phone numbermarora@iisc.ac.in ,+91-99808361743.Title of ProjectAI-enabled lung ultrasound imaging and sound analysis for triage and monitoring of COVID-194.Application for ICF ClusterAI/Data Science, Karnataka5.Co-Members of the Consortium *(Type – Indian Academia, Indian Business, International Academia, International Business, Indian non-commercial organisation, International non-commercial organisation, Other - specify)* Information not mandatory for the concept note stage.#OrganizationType1.National Institute for Advanced StudiesIndian Academia2.City, University of LondonInternational Academia3.St. John Medical College and Hospital/ St. Johns Research Institute BangaloreIndian Academia / Indian non-commercial organisation4.KPI Healthcare India Pvt. LtdIndian Business6.If you do not already have agreed consortium members, please indicate here the kind of partners you would like to hear from. Our consortium has 3 academic institutions, two from India, one from the UK, one medical institution and one Indian business thus we consider that the consortium is complete, but we might expand the consortium in future7. Are you content for us to publish your concept note on our web platform in order to encourage potential consortium members to reach out to you? Yes8.Funding Requested (up to ?250k) ~?230k9.Co-funding provided (if any) and sourceYesThe partners of the consortium will provide in-kind funding in terms of academic time, expert advice, access to commercial hardware for user trials and the research, access to computational resources such as GPUs and clusters to train the AI solutions as well as administrative and technical support.10.Please summarise your Innovation Challenge Fund research project in one sentence. 350 character limitDesign and build a teleoperated system to capture and analyse the sound waves and images provided by ultrasound scanner for triage and monitoring of COViD-19 patients remotely while keeping healthcare workers safe.11.Please describe your project further. What problem does your pilot seek to address? Who are the potential beneficiaries and other stakeholders? What are the inputs and activities, and what are the outputs? What does success look like after 12 months of funded research? How will your solution reach the market place? 2500 character limitRespiratory distress is the leading symptom and cause of mortality in COVID-19[1]. Though symptoms like dry cough, fever, reduced oxygen saturation and anosmia (loss of smell) are indicative symptoms of COVID-19, confirmation of the disease requires RT-PCR (real-time polymerase chain reaction) testing or imaging tests like computed tomography (CT) or X-rays. While RT-PCR test report only provides a positive or negative result – the impact of viral load on condition of the lung is not assessed. X-ray and CT provide alternative near real-time indications about the lung condition, however, both techniques have limitations. First, these techniques use ionising radiation, which is harmful and thus cannot be used repeatedly. Second, interpretation of radiographs is difficult and whilst CT scanners are not always available, due to elevated cost, especially outside teir-1 cities in India. Furthermore, in both cases, the preparation of patients for scan can expose healthworkers to COVID. Ultrasonic lung imaging biomarkers [2] have emerged as an alternative to imaging with ionising radiation to provide information about status of lung infection in real-time [3]. The information can be acquired in local manner by application of 12 zone method [4]. This approach currently has two limitations: first, the ultrasound probe needs to be operated by expert clinician, second, the interpretation of images can be challenging, even for a trained clinician. Complementary to ultrasound images, sound recordings can be obtained easily through digital stethoscopes and their ability to detect COVID-19 has already been demonstrated in initial studies [5]. Both ultrasound imaging and sound recordings provide signals that require processing to recognise and extract relevant information. Recent advances in image and audio signal processing using deep learning can enable effective enhancement and analysis for medical decision support. To enable capture of ultrasonic images and sound recording without significant exposure to health workers, we plan to integrate an ultrasound imaging probe and a digital stethoscope with a low-cost robotic arm, which can be operated from a safe distance. To ensure good quality of images and sound signals, the robotic arm will be operating in contact force control mode. Overall goal of this project is to develop an integrated system using existing hardware components (ultrasonic imager, digital stethoscope, tele-operated robotic arm) with novel AI-based software to effectively monitor hospitalized COVID patients in tier-2 tier-3 cities. With help of our consortium partner, KPI Healthcare with extensive experience of Indian healthcare market, we will be able to take this solution to healthcare providers in tier-2 tier-3 cities in India. In the future, the solution can also be exported to other countries, including UK.2443 Char.12.What is ‘technological’ about your proposed solution and why might this be appropriate to the challenge areas?1000 character limitOur proposed solution integrates available affordable hardware with data analytics tools to provide decision support during day to day management of COVID-19 patients in resource constrained settings.Image processing can improve image quality, extract quantifiable features, and help identify and localise anomalies. Computer vision methods have made significant progress in recent years, especially towards artificial intelligence and deep learning, which will be the emphasis of this project. Similarly, recordings of breathing and coughing sounds contain a great amount of information about the flow of air in the lungs and their internal condition. Again, recent improvements of sound processing with deep learning methods have shown significant success in sound signal enhancement and analysis in medical and specifically Covid-19-related tasks. By combining sounds and images in a deep learning solution, we will deliver high-quality monitoring of lung condition in a non-invasive, radiation-free, mobile, and inexpensive solution. By integrating the software with several hardware solutions and with the portfolio of the distributors, we can ensure availability in the market and promote uptake. 13.Is your proposed solution a response to the impacts of COVID-19, or an effort to contain the pandemic? If so, please explain. 1000 character limit Yes, our proposed solution is towards helping manage present CoVID pandemic. High quality monitoring of Covid-19 patients will save unnecessary journeys to major hospitals with CT facilities, where exposure to radiation as well as geography and cost are limiting factors, and ensure that patients with severe cases can get the treatment they need. Beyond CoVID, this research enhances capabilities of AI based analysis of ultrasonic imaging and sound recording of the lungs in general. Further teleoperated ultrasound system will be useful for high patient load setting like government hospitals where up to 400 patients might be undergoing scanning in a day. 14.Does your proposed solution contribute to combatting climate change or promoting a greener planet? If so, please explain. 1000 character limitReducing travel to larger hospitals may have a positive impact on CO2 emissions, but this is not the main goal of this project. 15.How is your proposal relevant to the development challenges of India? 1000 character limitThe approach proposed in this proposal is guided by the constrained resources of the Indian healthcare system. Firstly, unlike healthcare systems in wealthier countries, in India access to CT for imaging-based diagnosis is limited to major metropolitan cities only. Our approach to integrate affordable medical devices with AI based software tools will provide a safe, effective, and affordable system for management of COVID patients without need for CT scans on day to day basis. Technology developed for resource constrained setting, once proven effective, will provide cost savings in healthcare systems worldwide. 16.What consideration have you made of gender in developing your concept? Could your project address gender inequality or other kinds of inequality? 1000 character limitThe proposed work will be carried out under stated policy of gender equality of the host institutions. Furthermore, consortia will conform with equivalent policy from Innovate UK. Though the technology developed is gender agnostic and will benefit all genders equally it is worth noting that pregnant women are typically not allowed to undergo X-ray or CT scan. Since our proposed approach does not expose patients to these ionising radiations it ensures availability of treatment monitoring option for pregnant women as well. References:Li & Ma, Crit. Care 2020Guo, J Med Imaging 2018Convisar, Anesthesia & Analgesia 2020Doerschug, Annals Am. Thoracic Soc. 2013Brown, ACM KDD 2020 ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download