Employing Acoustic Features To Aid Neural Networks Towards Platform Agnostic Learning In Lung Ultrasound Imaging

Abstract

With the recent outbreak of COVID-19, ultrasound is fast becoming an inevitable diagnostic tool for regular and continuous monitoring of the lung. However, lung ultrasound (LUS) is unique in the perspective that, the artefacts created by acoustic wave propagation is aiding clinicians in diagnosis. In this work, a novel approach is presented to extract acoustic wave propagation driven features such as acoustic shadows, local phase-based feature symmetry, and integrated backscattering to automatically detect the pleura and to aid a pretrained neural network to classify the severity of lung infection based on the region below pleura. A detailed analysis of the proposed approach on LUS images over the infection to full recovery period of ten confirmed COVID-19 subjects across 400 videos shows an average five-fold crossvalidation accuracy, sensitivity, and specificity of 97%, 92%, and 98% respectively over randomly selected 5000 frames. The results and analysis show that, when the input dataset is limited and diverse as in the case of COVID-19 pandemic, an aided effort of combining acoustic propagation-based features along with the gray scale images, as proposed in this work, improves the performance of the neural network significantly even when tested against a completely new data acquisition.

Ultrasound-Based AI for COVID-19 Detection: A Comprehensive Review of Public and Private Lung Ultrasound Datasets and Studies

“…Durrani et al (2022) considered six distinctive scanning regions in their study. Panicker et al (2021) adopted the scan protocol of Soldati et al (2020) and also aimed at six acquisition points for data extraction. Quentin Muller et al (2020) scanned on ten thoracic sites in their study.…”

“…Variations in transducer types and frequency were also observed in the studies. For example, some studies used low-frequency (1-5 MHz) curved array (Chen et al, 2021;Huang et al, 2022;Camacho et al, 2022;Panicker et al, 2021) and phased array (Durrani et al, 2022;Arntfield et al, 2020) transducers. On the other hand, Roshankhah et al (2021) used both linear and convex transducers in multi-sites with a wide range of center frequencies.…”

Ultrasound-Based AI for COVID-19 Detection: A Comprehensive Review of Public and Private Lung Ultrasound Datasets and Studies

“…Durrani et al [35] considered six distinctive scanning regions in their study. Panicker et al [36] adopted the scan protocol of Soldati et al [24] and also aimed at six acquisition points for data extraction. Quentin Muller et al [37] scanned on ten thoracic sites in their study.…”

“…Variations in transducer types and frequency were also observed in the studies. For example, some studies used low-frequency (1-5 MHz) curved array [36,[39][40][41] and phased array [35,38] transducers. On the other hand, Roshankhah et al [45] used both linear and convex transducers in multi-sites with a wide range of center frequencies.…”

Review of Machine Learning in Lung Ultrasound in COVID-19 Pandemic

“…For multiscale severity based on B-lines, the model reached a weighted kappa of 0.65. Panicker et al presented a method to first detect the pleura via extracted acoustic wave propagation features; after obtaining the region below pleura, the infection severity was classified with VGG-16 from input regions [ 94 ]. They achieved an accuracy, sensitivity, and specificity of 0.97, 90.2, and 0.98, respectively, for 5000 video frames from ten patients over their infection to the full recovery phase.…”