Dear Colleagues ;
To find a way to improve the specificity of chest CT in COVID-19 we have conducted a project in Radiomics. Data gathering and Radiomic analysis are over. Now we need an AI/ML expert to analysis the results.
First of all we must find if any feature can be helpful for COVID-19 diagnosis.
If we find promising features we must select the most important ones ( for example by LASSO)
Then we must develop AI models to use these features and diagnosis the COVID ( for example SVM or NN, but we can try many different models).
If any AI experts likes to participate please send an email and CV to my address: [email protected]
If the results are positive then we try to publish the paper. If we can not find a way to improve CT specificity then we wont publish any paper from it.
The AI expert must be informed than there is no guarantee to publish a paper but if we can publish anything for sure we will add his/her name.
There is no fund in this project and all of us including me are working as volunteers.
Thank you very much
Houman Sotoudeh MD
Department of radiology, University of Alabama at Birmingham.
Awesome!
Recently I have worked in two technical startups that both focused on machine learning and computer vision. I can develop and train neural networks for both classification and image analysis.
I am ready to join your team to fight against COVID-19 by providing my time, expertise and computing/server resources available to me. I do not need any payment for that, I am after the virus.
Best
Audrius Meskauskas
Awesome!
Recently I have worked in two technical startups that both focused on machine learning and computer vision. I can develop and train neural networks for both classification and image analysis.
I am ready to join your team to fight against COVID-19 by providing my time, expertise and computing/server resources available to me. I do not need any payment for that, I am after the virus.
Best
Audrius Meskauskas
Hi,
i have some experience in AI Vision and biomedical imaging. Please take a look at:
- instance segmentation example
https://github.com/krzysztoffiok/c_elegans_fitness
- detection
https://github.com/krzysztoffiok/CNN-based-image-analysis-for-detection-and-quantification-of-neutrophil-extracellular-traps
- behavioral analysis
https://github.com/krzysztoffiok/CNN-behavioral-video-analysis-demo
I also can carry out Explainable AI (XAI) analysis for LM models (e.g. gradient boosting) with use of SHAP https://github.com/slundberg/shap which is probably the state-of-the-art in XAI for ML models. In brief, this method provides highest-quality feature importance analysis, also it allows to analyze sample-wise explanations of predictions.
Please contact me if you are interested in cooperation.
Best,
Chris
my suggestion is using CNN Algorithm, it's better than RNN, SVM
what is size of data set? which you are using and how many variables are there,?
pls, check https://pytorchlightning.github.io/lightning-Covid19 with in-progress development on Covid19 detection...
two things are here one that the technology is there for anything to do with and more for the scientific, one can use any basic classification algorithm for this, but the as told the second and most important is subject matter, who can compare with or with covid-19 , it was learnt that since it more of pulmonary and that extreme change will take place in very short period of time, so need is to analyse more test cases and so many cases you are dealing with right now
Sir
We can use deep learning algorithms like cnn for learning the disease pattern and predict the class label(disease category).dataset for training and testing is available with kaggle. Chest Xray and CT scan data is available
https://www.kaggle.com/praveengovi/coronahack-chest-xraydataset#Chest_xray_Corona_dataset_Summary.csv
CT scan
https://www.kaggle.com/andrewmvd/covid19-ct-scans
- Badges
- Science topic