PI or Contact
ICOVAI: Consortium on Covid-19 AI
Amsterdam UMC joined the consortium developing an AI tool to support overwhelmed medical staff with the reporting of CT scans.
AI image analysis
We aim to predict outcome and optimal treatment of patients with COVID-19 using machine learning from comprehensive clinical characteristics in a large cohort.
AI data analysis
COVID-19 CT Analysis
Collaborative hackaton between Amsterdam UMC and Radbound UMC on the develoopment of AI algorithms for analysis of chest CT for the detection of CORADS and CT severity score.
AI image analysis hackathon
Video AI for Social Distancing Recognition
In the 1.5-meter society it is crucial to maintain social distance. First and foremost, for the public health interest, but soon also to ensure viable operation of our society in times of COVID-19. This project studies whether video AI is able to recognize social distancing behavior in camera recordings, without the need to recognize individuals and any manual inspection of the recordings. The AI should analyze whether people are able to maintain social distance at work or in meeting places for large audiences, like airports, train stations, and shopping malls. The generated textual reports may inform relevant stakeholders to better organize our 1.5-meter society.
COVID-19 apps data sharing and privacy
The Systems and Networking Lab is applying its extensive research expertise on secure data sharing in Digital Data Marketplaces to the COVID-19 app space.
Data sharing infrastructures and model
Covid-19 Systematic Reviews
Healthcare professionals (e.g. nurses, physicians) as well as medical guideline developers and healthcare policymakers are expected to base their decisions on the latest available scientific evidence. To ensure that indeed the latest scientific knowledge is incorporated in their decisions, all up-to-date evidence and data needs to be found, extracted, appraised and combined or analyzed as soon as possible. To this date very little is known about the novel coronavirus, however tens of publications become available daily. The project is a collaboration between ILPS and Cochrane Netherlands aiming to automate systematic reviews around covid-19.
AI Information Retrieval, Information Extraction and Evidence Synthesis
Communication between health care providers and Deaf patients in times of COVID-19
It is very difficult for health care providers to communicate with their Deaf patients in times of COVID-19. Sign language interpreters are often not allowed to enter the hospital, and mouth caps make lipreading impossible. We will develop software that can automatically translate the most common questions (e.g., Where does it hurt?) and announcements (e.g., You have to stay at least one more night) into Dutch Sign Language. These translations will be displayed by means of pre-recorded videos if available, and otherwise by means of a sign language avatar which can produce sentences in sign language 'on the fly'.
Floris Roelofsen, Anika Smeijers