Mobile App for LabelLab
- Name: Udesh Kumarasinghe
- Email: mail@udesh.xyz
- University: Univesity of Colombo School of Computing,Sri Lanka
- GitHub Profile: https://github.com/UdeshUK
- LinkedIn Profile: https://www.linkedin.com/in/udeshuk/
LabelLab Mobile App focuses on creating a mobile application to classify and label all kind of images without being specific to a category to satisfy more use cases. LabelLab mobile app’s motivation is to allow the user to use a classification model in the field using their handheld device.
For the GSoC 2019, I've initiated the LabelLab Mobile app project from scratch and implemented the required functionality using Flutter. Also, I've collaboratively implemented modules of LabelLab server which are required for LabelLab Mobile app's functionality. I have successfully implemented image classification from camera or gallery, classification history, admin features, authentication with social logins and documented information needed to set up the project.
- Collaboratively designed and finalized labellab-server endpoints
- Designed mobile app's UI flow and wireframes
- Initialized flutter project and configured dependencies
- Implemented sign in and sign up frow and integrated with the labellab-server
- Integrated Google Sign In and GitHub SSO
- Created the user profile and update profile picture edit feature
- Created classification endpoints in the labellab-server to be used for classifications
- Implemented image classification flow to classify images using camera or gallery
- Created history section to list out user's previous classifications
- Implemented project management(create, edit, delete) features
- Implemented user search endpoint and functionality to search users
- Implemented project's user management (add or remove members)
- Implemented project's label management (create, edit, remove)
- Implemented single and batch image upload with camera or gallery
- Implemented uploaded image management features
- Created a labeling tool to add labels to uploaded images using rectangular or polygonal selections
- Wrote tests for implemented endpoints of labellab-server
- Implement path tracking feature by integrating with the labellab server