|
Sounds of CovidDuring the Covid pandemic we has a very fruitful and exciting collaboration with Cecilia Mascolo and her team in computer science. Thanks to a very energetic campaign, we crowdsourced an enormous database of sounds of people coughing, speaking and breathing, with and without Covid. This was then used to train ML models. A number of papers have been published on this work, which has received a lot of attention. The first papers focused on yes/no diagnosis, whereas later papers could investigate longitudinal (time series) data of the same person, showing the progression of sound in and out of the diseased state. See for example [J.Han, et al., Journal of Medical Internet Research 25 e44804 2023] ML in image analysisWe are of course users of ML in the many useful packages that now exist for cell segmentation. We were also involved in very interesting work with Neill Campbell in Bath, see for example [J.Goodier, et al., Frontiers of Medical Technology, 123-129 2024] developing new methods with the aim of providing a robust estimate of uncertainty. The test case was detection of red blood cells infected by malaria, on which a large GCRF project hled by Richard Bowman had got us together and connected to clinicians in Tanzania, leading to a very large curated collection of malaria smear image samples.
| |
 |
 |
 |
