Automation of lung ultrasound imaging and image processing for bedside diagnostic examinations

The causes of acute respiratory failure can be difficult to identify for physicians. Experts can differentiate these causes using bedside lung ultrasound, but lung ultrasound has a considerable learning curve. We investigate if an automated decision-support system could help novices interpret lung ultrasound scans. The system utilizes medical ultrasound, data processing, and a neural network implementation to achieve this goal. The article details the steps taken in the data preparation, and the implementation of the neural network. The best model’s accuracy and error rate are presented, along with examples of its predictions. The paper concludes with an evaluation of the results, identification of limitations, and suggestions for future improvements.

[1]  S. Speidel,et al.  Robot-Assisted Minimally Invasive Surgery—Surgical Robotics in the Data Age , 2022, Proceedings of the IEEE.

[2]  Edmund Y. Lam,et al.  Transfer Learning U-Net Deep Learning for Lung Ultrasound Segmentation , 2021, ArXiv.

[3]  L. Demi,et al.  Deep learning applied to lung ultrasound videos for scoring COVID-19 patients: A multicenter study. , 2021, The Journal of the Acoustical Society of America.

[4]  Jérémie F. Cohen,et al.  THORACIC IMAGING TESTS FOR THE DIAGNOSIS OF COVID-19 , 2020, The Cochrane database of systematic reviews.

[5]  Zhiyong Lu,et al.  Automated abnormality classification of chest radiographs using deep convolutional neural networks , 2020, npj Digital Medicine.

[6]  Cristiano Saltori,et al.  Deep Learning for Classification and Localization of COVID-19 Markers in Point-of-Care Lung Ultrasound , 2020, IEEE Transactions on Medical Imaging.

[7]  Tamas Haidegger,et al.  Autonomy for Surgical Robots: Concepts and Paradigms , 2019, IEEE Transactions on Medical Robotics and Bionics.

[8]  Thomas Brox,et al.  U-Net: Convolutional Networks for Biomedical Image Segmentation , 2015, MICCAI.

[9]  Andrew Zisserman,et al.  Very Deep Convolutional Networks for Large-Scale Image Recognition , 2014, ICLR.

[10]  D. Lichtenstein.  Lung ultrasound in the critically ill , 2014, Annals of Intensive Care.

[11]  Milan Sonka,et al.  3D Slicer as an image computing platform for the Quantitative Imaging Network. , 2012, Magnetic resonance imaging.

[12]  Geetha Muthusamy,et al.  Deep Learning-based Transfer Learning Model in Diagnosis of Diseases with Brain Magnetic Resonance Imaging , 2022, Acta Polytechnica Hungarica.

[13]  I. Rudas,et al.  Robotics and Intelligent Systems Against a Pandemic , 2021, Acta Polytechnica Hungarica.

[14]  Handbook of Medical Image Computing and Computer Assisted Intervention , 2020 .

[15]  D. Lichtenstein.  Basic Applications of Lung Ultrasound in the Critically Ill: 2 – The Ultrasound Approach of an Acute Respiratory Failure: The BLUE Protocol , 2010 .