https://github.com/rekalantar/covid19-detector-ios-application

Covid-19 Detector iOS Application Developed by Reza Kalantar
https://github.com/rekalantar/covid19-detector-ios-application

classification covid-19 deep-learning detection ios tensorflow transfer-learning

Last synced: 8 days ago
JSON representation

Covid-19 Detector iOS Application Developed by Reza Kalantar

• Host: GitHub
• URL: https://github.com/rekalantar/covid19-detector-ios-application
• Owner: rekalantar
• License: mit
• Created: 2020-07-23T10:58:07.000Z (over 4 years ago)
• Default Branch: master
• Last Pushed: 2022-11-14T09:57:37.000Z (about 2 years ago)
• Last Synced: 2024-11-09T05:15:25.162Z (2 months ago)
• Topics: classification, covid-19, deep-learning, detection, ios, tensorflow, transfer-learning
• Language: Swift
• Homepage:
• Size: 60.6 MB
• Stars: 2
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# Covid-19 Detector iOS Application

 Covid-19 Detector iOS application with embedded deep learning classifier for detecting Covid-19, viral and bacterial pneumonia from frontal chest X-ray scans.

![alt text](https://raw.githubusercontent.com/rekalantar/Covid19-Detector-iOS-Application/master/application_demo.png)

### Software requirements:

Xcode/

Cocoapods/

Firebase package/

TensorFlowLite package

### Instructions:

(1) Clone directory and insert command 'pod install' in project root directory from the terminal

(2) Open the 'CovidCXR.xcworkspace' folder and run the app through the iPhone/iPad simulator

(3) Click 'Load' from the homepage to upload a photo from gallery and click 'detect' for inference

(4) To take photos for detection from device directly, an external iOS device needs to be connected and run instead of the simulator

(5) The 'Draw' button on the image can be activated to outline regions-of-interest on images

(5) The X-ray images loaded in the application can be uploaded to a database for crowd-sourcing. Note: for patient security all texts from scans are automatically removed before uploading