github-actions[bot]
commented
5 days ago Thank you for creating this issue! đ We'll look into it as soon as possible. In the meantime, please make sure to provide all the necessary details and context. If you have any questions reach out to LinkedIn. Your contributions are highly appreciated! đ
Note: This repo is for beginners to learn and start with Opensource we won't accept more than 10 issues from a single person, This restriction applies to Gssoc project which has a similar kind of adding folder files, Points will be reduced when we find Spam.
I Maintain the repo issue twice a day, or ideally 1 day, If your issue goes stale for more than one day you can tag and comment on this same issue.
You can also check our CONTRIBUTING.md for guidelines on contributing to this project.
ashis2004
Is there an existing issue for this?
Feature Description
Built and trained a convolutional neural network in Keras with TensorFlow as backend from scratch to predict patients if they were infected with COVID-19 or not using their chest X-ray images. Matplotlib was used for data visualization. Data preprocessing and data augmentation was carried out using tensorflow 2.0 The model used was sequential with a combination of convolutional layers, pooling layers, dropout layers, dense layers with relu activation and output layer with sigmoid activation. The dataset contained the lungs X-ray images of both groups i.e non-covid and covid infected patients. The dataset was obtained from kaggle. Metrics chosen for model evaluation were Training set, test set and validation set accuracy. Adam optimizer with learning rate of 0.001 was choosed for gradient descent The entire project was carried out on the Google Colab environment Results of the model were following:
Training Set Accuracy : 98.41 %
Training Set Loss : 0.0490
Validation Set Accuracy : 97.51 %
Validation Set Loss: 0.0759
Test Set Accuracy : 94.83 %
Test Set Loss : 0.1141
Use Case
The prediction of COVID-19 from chest X-ray images using Convolutional Neural Networks (CNNs) has several critical use cases in the healthcare sector. Firstly, it enhances diagnostic accuracy and speed in clinical settings, especially in regions with limited access to PCR testing facilities. By rapidly analyzing X-ray images, CNNs can assist radiologists in identifying COVID-19-related abnormalities, enabling timely isolation and treatment of infected individuals. Secondly, CNN-based models can serve as valuable tools in emergency departments, where quick decision-making is essential. They help in triaging patients, determining the severity of lung involvement, and prioritizing those needing immediate medical attention.
Benefits
No response
Add ScreenShots
Priority
High
Record