logo SBA

ETD

Digital archive of theses discussed at the University of Pisa

 

Thesis etd-02012021-130738


Thesis type
Tesi di laurea magistrale
Author
SANTO, RUGGIERO
URN
etd-02012021-130738
Thesis title
Dynamic neural networks for COVID-19 severity prediction from lung ultrasound
Department
INFORMATICA
Course of study
INFORMATICA
Supervisors
relatore Bacciu, Davide
correlatore Faita, Francesco
Keywords
  • classification
  • COVID
  • COVID-19
  • LUS
  • Neural Networks
  • ultrasound
  • video
Graduation session start date
05/03/2021
Availability
None
Summary
The thesis dealt with preparation and predictive analysis of lung ultrasound video with the aim of predicting severity of COVID-19 effects. The acquired videos were cleaned, organised and processed for training and benchmarking different deep learning models from the action recognition literature. The developed models were also evaluated from a clinical perspective against medical expert opinion.
File