Skip to content

dhruv-varshney/COVID-19-Testing-Using-X-Ray-Images-Web-App-Development

Repository files navigation

PROBLEM STATEMENT :

Existing process of the getting information of covid-19 x-ray images and Pneumonia normal images is extracted from online available resources. World Health Organization has declared Covid-19 as a pandemic. There are a number of test kits available but are either expensive or take time to detect the virus. Deep Learning is State-of-the-art machine learning algorithm and has shown massive development in detecting Covid-19 with the help of chest X-rays. In this blog, we will learn to detect Covid-19 with help Chest X-rays using CNN architechture.Thus providing this purpose with necessary functionalities.

SOFTWARE REQUIREMENTS SPECIFICATIONS :

Software requirements : Minimum software requirements are: 1)Tool : Anaconda, CNN Architechture 2)Operating System : Windows XP/7,8,10 3)Scripting Language : Python, Flasks 4)X-Ray Images : COVID-19, Pneumonia normal

System Design

Our proposed deep learning-based COVID-19 detection comprises several phases, as illustrated in Figure. The phases are summarised in the following five steps: Step 1: Collect the chest X-ray images for the dataset from COVID-19 patients and healthy persons. Step 2: Generate Chest X-ray images using data augmentation. Step 3: Represent the images in a feature space and apply deep learning. Step 4: Split the dataset into two sets: a training set and a validation set. Step 5: Evaluate the performance of the detector on the validation dataset.

DATASET :

COVID-19 images used in training are from https://github.com/ieee8023/covid-chestxray-dataset

Normal Images

Dataset: Chest X-Ray Images (Pneumonia), https://www.kaggle.com/paultimothymooney/chest-xray-pneumonia It has two types of X-Ray images Normal and Pneumonia but only normal X-Rays have been used from this dataset so that Model learns to identify Covid-19 cases.

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published