The COVID-19 pandemic caused by SARS-CoV-2 remains a significant issue for global health, economics, and society. The person-to-person transmission of the virus led to the global pandemic and self-isolation of people. No way of destroying the virus has been found since its discovery. Its global spread has caused lakhs of deaths. The disease has caused a major impact on the economy, the education system, and physical and mental health. The current short review summarizes the available knowledge about pandemic, morphology, and characteristics of coronavirus, diagnosis, drugs being used for treatment . Our model can be employed to assist radiologists in validating their initial screening, and can also be employed to immediately screen patients.
What it does
The proposed model has an end-to-end architecture without using any feature extraction methods, and it requires raw chest X-ray images to return the diagnosis. A Flask App is built to which an X-Ray image has to be given as input. It does predict whether the given X-Ray is of a covid infectious person or not.
How we built it
Built with TensorFlow Keras, Flask. The deep learning model is made with Convolution Neural Networks VGG16 Architecture.VGG16 is loaded with pre-trained weights for imagenet. After defining the model, we need to load the input image with the size expected by the model, in this case, 224×224. The pixel values then need to be scaled appropriately for the VGG model. A flask app is built for user interaction to use the model for prediction of x rays as covid positive or negative
Challenges we ran into
I was facing challenges when I was making a flask app for the model. There were TensorFlow installation errors earlier and thereafter I had some h5py library errors. I solved multiple of them by uninstalling and downgrading the versions.
Accomplishments that we're proud of
This model itself will help during the ongoing pandemic to detect the possibility of covid infection within seconds via X-ray images. Thus helping us to combat the ongoing pandemic.
What we learned
This project helped me understand CNN Architecture and data augumentation.
What's next for Covid-19 detection using Deep Learning
I would be deploying my model as soon as possible. And would like to extend it a bit like adding a covid chatbot to it.