CovCT: An Android Application for COVID-19 & Non-COVID-19 Detection/Classification using Lung Computed Tomography Images Paper Link
Student : Aryan Verma
Mentor : Monjoy Saha
Chest CT scans has a potential role in the diagnosis and prognostication of COVID-19 and the designing of a portable and cost-effective diagnosis system capable of running on resource constrained devices like Android OS Phones would indeed improve the clinical utility of chest CT.
CovCT is an android application developed as a GSoC 2021 project to detect COVID-19 (CoronaVirus Disease 2019) from Chest CT Scan images. The application uses the power of Deep Neural Networks to detect COVID-19 from Chest CT Scans and further generate a heatmap which shows the affected regions in the chest.
- Segmentation of Lung region from the CT scans through a robust algorithm of Computer Vision.
- Predicting the CovThe Video Demonstration for app in use can be seenid-19 percentage in the Lungs from the CT Scan.
- Generating the heatmaps indicating the regions of Covid-19 infection in the lungs.
- Augmenting the heatmaps with Segmented lungs and saving.
The Video Demonstration for app in use can be seen here
- The app can be installed by directly downloading the Android Application Package (APK) File located at : "Demonstration and results/app-debug.apk" into the Mobile phone device with Android OS.
- For building the project on your local machine, clone the repo and Open it with Android Studio.
- After installing or building the app, the first screen which is encountered is a Splash screen with animation and CovCT Logo. Before this step remember to download the sample images from the "Demonstration and Results/Image Results" folder for testing purposes. The second screen comes for asking permission from user of :
- WRITE_EXTERNAL_STORAGE
- READ_EXTERNAL_STORAGE
- The third screen Renders Home Page of the application which shows option of Scrolling to the next fragment which is CT Scan loading Fragment.
This Activity also contain Splits navigating to the extra options for sharing app, developer info and link to this github repo.
- The fifth screen gives us the option to Load our CT Image from Phone Storage. This acitivity checks for the size of image and type of image and throws error if images are of unequal dimensions. Following the Uploading of the image comes the further steps in the same activity :
- When user Press "Predict" Button the image is sent to Segmentation Algorithm which segments the lungs from the image. After segmentation of the lungs, the deep learning inference is generated and diplayed on the screen.
- The last screen appears after the inference resulted in more than 50% covid-19 prediction. Now the image is passed to Heatmap Generation Algorithm which is ScoreCAM. This step generates the heatmap and augments it on the lungs, after which Using the trackbar and checkbox, heatmap gradient colour and mask of lungs can be adjusted respectively. Further this modified and augmented image can be saved by one click.
- Usage of android studio to build the app with java and XML as backend languages.
- OpenCV Library for Lungs Segmentation with custom Segmentation Algorithm and other Image Processing Tasks.
- Nd4j Scientific Computing Library for JVM along with OpenCV JAVA for Heavy Image processing Tasks and handling N-Dimensional arrays while generating Heatmaps.
- Keras library with Tensorflow backend for the creation of deep learning model to detect Covid-19 in Chest CT Images.
- Python Scripting on Google Colab/Jupyter Notebook for the model to be trained. And integration of the .tflite file which reduces the size of AI model and hence the app.
- Numpy and Pandas Libraries for ease of Model Training
