Streamlit - Annotator

Streamlit Annotator provides an easy-to-use interface for annotating your data, inspired by streamlit-drawable-canvas

Features

• Easy-to-use interface for annotating images
• Drawing bounding boxes
• Rotating, undoing, redoing, or deleting bounding boxes
• Saving annotation data in canvas JSON format and LabelMe JSON format

Demo

Draw bounding box

Remove bounding box

To remove a bounding box, switch to transform mode and double-click on the box you want to delete.

Installation

Create a new conda environment and install the required dependencies:

conda create -n streamlit-annotator python=3.10 -y
conda activate streamlit-annotator
pip install requirements.txt

Usage

Run the Streamlit app:

streamlit run app.py

This will launch the Streamlit Annotator interface in your default web browser.