Quantification of pulmonary involvement in COVID-19 pneumonia by means of a cascade of two U-nets: training and assessment on multiple datasets using different annotation criteria.
Francesca LizziAbramo AgostiFrancesca BreroRaffaella Fiamma CabiniMaria Evelina FantacciSilvia FiginiAlessandro LascialfariFrancesco LaruinaPiernicola OlivaStefano PifferIan PostumaLisa RinaldiCinzia TalamontiAlessandra ReticoPublished in: International journal of computer assisted radiology and surgery (2021)
We analysed the impact of using data samples with different annotation criteria in training an AI-based quantification system for pulmonary involvement in COVID-19 pneumonia. In terms of vDSC measures, the U-net segmentation strongly depends on the quality of the lesion annotations. Nevertheless, the CT-SS can be accurately predicted on independent test sets, demonstrating the satisfactory generalization ability of the LungQuant.
Keyphrases
- coronavirus disease
- sars cov
- pulmonary hypertension
- rna seq
- virtual reality
- deep learning
- computed tomography
- artificial intelligence
- convolutional neural network
- respiratory syndrome coronavirus
- big data
- respiratory failure
- contrast enhanced
- community acquired pneumonia
- single cell
- image quality
- magnetic resonance imaging
- magnetic resonance
- positron emission tomography
- pet ct