Supervised machine learning (ML) projects rely heavily on annotated datasets. Tools can be developed to facilitate and make the task of annotation / labeling more efficient and scalable.

This project consists of a simple android app that can be used to label images into binary classes. The app is built using React Native.

An implementation for an image server API is also available (image_server/image-server.py) from which the app will connect to to serve images that need to be annotated.

The server requires the following modules:

• flask
• flask_restful
• pandas

Steps needed during development:

• Run image-server.py <path-to-data>.
• Run python -m SimpleHTTPServer 8800 in the image directory (<path-to-data>/img).
• Run npm start in the project root directory.
• Run an Android Emulator or connect an android device.
• Open another terminal in the project root directory and run react-native run-android. Alternatively, Android Studio can be used to build and install the app to the emulator or device.

Screenshots for car labeler: