Deep learning models for covid-19 chest x-ray classification: preventing shortcut learning using feature disentanglement

HIGHLIGHTS

• who: Anusua Trivedi and collaborators from the Systems, Pensacola, FL, United States of America, Department of Ophthalmology, of have published the research: Deep learning models for COVID-19 chest x-ray classification: Preventing shortcut learning using feature disentanglement, in the Journal: PLOS ONE of 10/08/2021
• what: The authors propose adding feature disentanglement to the training process. In the experiments the authors used the recent SOTA lung VAE model to create lung masks and implementation of histogram equalization from OpenCV . This model has been trained on CXR imagery, collected before the COVID-19 pandemic . . .
  

   

  

  If you want to have access to all the content you need to log in!

  Thanks :)

  Username or Email

  Password

  Remember me

  Lost your password?

  If you don't have an account, you can create one here.