Logo
Logo

Research

1-6 of 11 publications

Self-supervised pretraining improves performance and inference efficiency in multiple lung ultrasound interpretation tasks

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.

Research
Ultrasound

Prospective real-time validation of a lung ultrasound deep learning model in the ICU

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.

Ultrasound
Research

Introducing medical students to deep learning through image labelling: a new approach to meet calls for greater artificial intelligence fluency among medical trainees

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.

Ultrasound
Education
Research

Improving the generalizability and performance of an ultrasound deep learning model using limited multicenter data for lung sliding artifact identification

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.

Ultrasound
Research

Exploring the utility of self-supervised pretraining strategies for the detection of absent lung sliding in M-mode lung ultrasound

This study explores the efficacy of self-supervised pretraining in improving performance for lung sliding classification in M-mode lung ultrasound images, highlighting its superiority over full supervision.

Ultrasound
Research

Enhancing annotation efficiency with machine learning: automated partitioning of a lung ultrasound dataset by view

We propose a method leveraging hierarchical organization of annotating tasks to increase efficiency in annotating lung ultrasound images, demonstrating significant time savings and higher throughput in view-specific annotation tasks.

Ultrasound
Research
Logo
general@deepbreathe.ai

© 2024, DEEP BREATHE, INC