It depends on what you consider new ideas. New ideas on classification or Covid 19? One thing I would suggest is gather data (image and non image data) of Covid 19 patients, alive or dead and test unsupervised machine learning methods in order to find patterns of the patients and determine the probability of survival. This is the very first idea that has crossed my mind. Also, read the current literature on Covid 19 and machine learning approaches that have been employed on the former. You can come up with new ideas. It may be the beginning for a Master or even Ph.D

I was just about to suggest what Constantinos Spanakis replied.

If you have data on the patients, such asl; what happened to them after the X-rays imaging, any physiological data like background diseases, age, and so on, you could build a prognosis tool that will help predict the course of the disease in new patients.

Plugin the X-ray and any available physiological data, and get a prediction of likelihood survival and allocate treatment accordingly

In my view, you should start by collecting the real medical data of COVID 19 patients with their pre-medical conditions. Maybe this will help you to make a connection between the previous diseases and their effect on a COVID 19 patient. You can develop a Deep learning scheme based on these data sets, which may be extremely helpful in the future for a COVID patient to analyze their vulnerability.

Md. Muhtadir Rahman

If you couple a CNN with dense layers at the end, you can do regression tasks, for example predicting the age of the patient from the CT image. Also, using auto-encoder CNNs you are able to segment medical images (check U-net).

Hello,

I was working on brain MRI scans for predicting dyslexia where we tried to find out the voxels which are contributing for dyslexia. You can try to find the interesting region leading for your classification. That will also help the literature for focusing on those areas. Try GLM analysis something similar to nilearn.

You can diagnose if the patient suffers from other diseases or if he needs ICU. This application is very important for insurance and healthcare companies.

Thanks all.

this link may be help you

Chapter Medical Image Classification Using Deep Learning