Mercedes Torres Torres and Joy Egede are postdoctoral graduates of the School of Computer Science. Mercedes is a Transitional Assistant Professor after graduating from the Horizon CDT, and Joy is a Research Fellow who completed her PhD within the International Doctoral Innovation Centre. Both are now based within the Computer Vision Lab, carrying out multidisciplinary research within a medicine and healthcare context, using innovative machine learning, image analysis and processing techniques.
Two examples of research projects that Mercedes and Joy have collaborated on with experts from the neonatal medical profession are provided below:
Babyface: Gestational age estimation tool
The Gestational Age Estimation project was funded by the Bill and Melinda Gates Foundation and brought together an interdisciplinary team of researchers in the Schools of Medicine and Computer Science.The challenge was to develop a tool to accurately calculate the gestational age of newborn babies using digital imaging, with the idea of deploying said tool in countries that have little or no access to ultrasound scanners and trained personnel. We developed state-of-the-art deep learning algorithms that were able to accurately estimate the gestational age of newborns within days.Within this project, the Babyface crowdsourcing smart phone app was developed, which was used by parents and guardians to upload photographs and anthropometric measures of their babies in order to collect new securely-stored data to help doctors estimate whether a baby was born prematurely.Due to its potential international real-world impact on society, the project received coverage in the national and international press, including an article in The Guardian and an interview on BBC Radio Nottingham
Neonatal pain assessment tool
Joy had already completed her PhD research on automatic pain assessment for adults and neonates using face video, and in 2018 successfully secured internal impact funding from the University of Nottingham to work on developing new video-based pain assessment tools for newborns, in collaboration with the National Hospital in Nigeria.This impact acceleration project built upon the progressive pain assessment techniques that emerged from Joy’s doctoral research, and resulted in:
- the development of a new automated newborn pain assessment tool which produced results that were comparable to manual assessment by experienced Neonatal ICU nurses
- the creation of a new 11-point Neonatal Face and Limb Acute Pain Scale (NFLAPS), specifically suited for computer vision technologies
- the development of a new Acute Pain in Neonates database containing over 250 video recordings of various medical procedures and NFLAP pain annotations (scored by NICU nurses) for each video frame
Going forward the database will assist computer vision researchers in developing more efficient newborn pain assessment tools, consequently leading to improved infant care. Work continues to improve the performance of the model. Joy is also an award-winning prize recipient from the 2020 L’Oréal-UNESCO For Women in Science UK & Ireland Rising Talent Programme.