Lung Ultrasound Imaging and Image Processing with Artificial Intelligence Methods for Bedside Diagnostic Examinations
Gábor Orosz, Róbert Zsolt Szabó, Tamás Ungi, Colton Barr, Chris Yeung, Gábor Fichtinger, János Gál, Tamás Haidegger
Abstract
Artificial Intelligence-assisted radiology has shown to offer significant benefits in clinical care.Physicians often face challenges in identifying the underlying causes of acute respiratory failure.One method employed by experts is the utilization of bedside lung ultrasound, although it has a significant learning curve.In our study, we explore the potential of a Machine Learning-based automated decision-support system to assist inexperienced practitioners in interpreting lung ultrasound scans.This system incorporates medical ultrasound, advanced data processing techniques, and a neural network implementation to achieve its objective.The article provides a comprehensive overview of the steps involved in data preparation and the implementation of the neural network.The accuracy and error rate of the most effective model are presented, accompanied by illustrative examples of their predictions.Furthermore, the paper concludes with an evaluation of the results, identification of limitations, and recommendations for future enhancements.