The repository documents code required to process the data, train the VGG-based and its variant models and test the ultrasound image classifier to verify the performance.
The dataset is curtsy of Dr. Marini of URMC and is currently proprietary. Hence, it is unavailable openly.
healthcare has been increased in importance. Many patients hailing from developing countries have limited access to adequate health care. Using portable and more accessible option of using lung ultrasound(LUS) volume sweep imaging for initial test and prognosis is a potential solution
We developed methods for the automatic detection of Pneumonia from Lung Ultrasound (LUS) recordings. Our results show that one can accurately distinguish LUS samples from patients with respitory illness from healthy controls.
As mentioned earlier the lung ultrasound images are not released for open source as of now.
Find all details on how to reproduce our experiments and train your models on ultrasound data in the pocovidnet folder.
- We are using the part of the method and open source code cited in POCOVID-Net preprint
- In case of any problem with the code, please open an issue.