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 - Takeaways - MDSpire

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

  • By

  • Francesca Lizzi

  • Abramo Agosti

  • Francesca Brero

  • Raffaella Fiamma Cabini

  • Maria Evelina Fantacci

  • Silvia Figini

  • Alessandro Lascialfari

  • Francesco Laruina

  • Piernicola Oliva

  • Stefano Piffer

  • Ian Postuma

  • Lisa Rinaldi

  • Cinzia Talamonti

  • Alessandra Retico

  • October 26, 2021

  • 0 min

Share

  • 1

    The LungQuant system utilizes a cascade of two U-nets for fully automated segmentation of lungs and COVID-19 lesions from CT scans.

  • 2

    The study emphasizes the impact of dataset variability and annotation criteria on the performance of deep learning segmentation models.

  • 3

    Five publicly available datasets were used to train and evaluate the LungQuant system, highlighting the need for diverse training samples.

  • 4

    The system outputs lung volume affected by COVID-19 lesions and assigns a CT severity score based on the percentage of lung involvement.

  • 5

    The U-net architecture employed in LungQuant includes a compression path for feature extraction and a decompression path for output generation.

Original Source(s)

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

  • by Francesca Lizzi, Abramo Agosti, Francesca Brero, Raffaella Fiamma Cabini, Maria Evelina Fantacci, Silvia Figini, Alessandro Lascialfari, Francesco Laruina, Piernicola Oliva, Stefano Piffer, Ian Postuma, Lisa Rinaldi, Cinzia Talamonti, Alessandra Retico

  • October 26, 2021

Related Content

MIS-C: Diagnosis, Management, and Outcomes

Hybrid Care Modifications in the Delivery of Nonpandemic Care During the COVID-19 Pandemic: Scoping Review

AI in Surgery: Debate Highlights Benefits, Gaps

Surgeons and researchers clash over whether current AI systems are ready for clinical use — and what responsible adoption would require.