Discontinued service
email dlinradiology@gmail.com if you need to get in touch
Automatic COVID-19 burden load evaluation
This is an automatic process that on a server evaluates the disease burden of COVID19 changes on axial CT scans. A deep learning model annotates/segments all slices and calculates the volumes of lungs, ground-glass, and consolidation. The results are sent back as a report in PDF format. The report also includes 6 example slices with annotations for quality control.
To use, first select the individual anonymized DICOM files with the button below. If this goes well, then the submit button will activate and turn green. After pressing the submit button, the image data will be sent to a server, processed, and the resulting report will be sent back for you to download. With good internet, a 300 slice CT should take less than 5 minutes in total (most of which is the uploading of image data). Tested on Chrome, but should work with most other browsers too.
If you want to apply the model locally in the browser, then you can use the model that is available in MedSeg.
Troubleshooting:
The last update of models was on the 23rd of April.
One error that we are seeing is incorrect volumetric calculation in a few cases where we think the problem is different slice thickness from slice distance.
If you find errors or want to get in touch for other reasons, please don’t hesitate to contact us at dlinradiology@gmail.com. There could be an error due to simultaneous use by several people at the same time. If you get an error, you could try once a little later to see if that was the problem.
Acknowledgments would be greatly appreciated if you find value in this. Also, you can suggest what you would like to see in our next automatic segmentation report.