1-6 of 12 publications
This study aims to develop and validate a deep learning model for detecting pleural effusions in lung ultrasound images, with adaptable performance characteristics tailored to specific clinical scenarios.
In this study, we investigated the effect of utilizing proximal, distinct images from the same B-mode ultrasound video as pairs for SSL.
In this study, we investigated whether self-supervised pretraining could produce a neural network feature extractor applicable to multiple classification tasks in B-mode lung ultrasound analysis.
This prospective study assessed the accuracy of a deep learning model deployed in real-time at the bedside to differentiate between normal and abnormal lung parenchyma on lung ultrasound in critically ill patients.
Our approach bridges the gap between medical education and AI by empowering medical students to excel in medical imaging and labeling, fostering competence while generating valuable data for AI integration in clinical practice.
We report a multicenter study that exploits limited available external data to improve the generalizability and performance of our lung sliding model while identifying poorly performing subgroups to inform future iterative improvements.
© 2024, DEEP BREATHE, INC
