LS-MinIO-CV-Setup

A project for processing data and building project for label studio for Computer Vision tasks.

• LS-MinIO-CV-Setup
  • Test Environments
  • Commands
  • Data Folder Structure
    • Image Classification
  • TODO
    • Setups
    • Evaluation
    • Export

Test Environments

• Label Studio: 1.8.0

Commands

• setup
  • generate .env & docker-compose.*.yml
  • print command
• upload
  • root path
• create
  • project name
  • ...

Data Folder Structure

Image Classification

TODO

Setups

• Upload data to MinIO
• Create json files for data
• Create Project via Python SDK
• Adjust Tab View

Evaluation

• Validate task count with s3 storage (to avoid annotator deleting task)
• Validate tasks all annotated
• Evaluate annotator performance
• Evaluate annotation quality
• Make workflow more generalized

Export

• Export annotations
• Convert annotations to common CV task format