This app allows users to train models to predict clinical condition (e.g., Covid-19), or patient outcome (e.g., hospitalization), based on Chest X-Ray images. The user provides the images they want to train the model on, along with labels of the actual classification for each image. Once trained the model can then be applied to new datasets. Further details on how this app works can be found here. 

Files should be provided in either .png or .dcm format, with multiple files provided in a .zip compressed folder. We recommend that at least 200 images should be used for training new models. Using more images for training will likely improve results. Labels should also be provided in .csv format: with 1 indicating the presence of a condition, and 0 indicating the control group. If multiple labels are provided instead, in string format, in a single column, then a multilabel model will be trained instead.

Limitations:

Should be used for research purposes only.

This is not a clinically validated tool. Do not use CXR for self-diagnosis and seek help from your local health authorities.

Results may not generalize well to other patient populations or manufacturers not used in training.

Risks: Although neither Google nor Superbio permanently store any data processed by these models, it is the data owner's responsibility to ensure that Personally identifiable information (PII) and Protected Health Information (PHI) are removed prior to data upload.