Sam Labeler

Sam Labeler is a labeling tool for image data utilizing Segment Anything Model (SAM) version 1. This tool provides an intuitive interface for users to easily label images for machine learning tasks, particularly in segmentation tasks.

Table of Contents

• Upcoming Features
• Documentation
• Code Refactoring
• Completed Tasks

Installation

To get started with Sam Labeler, follow these steps:

1. Clone the repository:

   git clone https://github.com/AlbertZhaoCA/sam-labeler.git
   cd frontend
   pip install -r requirements.txt

   We recommend using a virtual environment to manage dependencies like anaconda or venv.

Upcoming Features

• Feature 1: Add React Frontend for click labeling
  • Next.js frontend
  • Add click labeling
  • Finalize data fetch pattern and its implementation in Next.js.
• Feature 2: Use sqlite or postgres for to be labeled images information storage and management
  • sqlite and sqlalchemy are used for storing labeled images
    • Decide whether to migrate to postgres Explore M3 or other storage services for efficient cloud storage (in the upcoming version 2.0).

Documentation

• Write comment
  • main.py
• Create a troubleshooting guide
• Complete the README.md file

Code Refactoring

• Refactor utility functions

Completed Tasks

• Frontend prototype
• Backend prototype