Overview

Django Labeller is an innovative tool designed for Python developers looking to create segmentation datasets effortlessly. Lightweight and versatile, it supports integration with popular frameworks such as Django, Flask, and Qt, making it accessible for various development environments. Whether you are building a web app or a desktop application, Django Labeller provides the necessary functionality to efficiently annotate images for deep learning projects.

The uniqueness of this tool lies in its support for multiple annotation types, including polygons, boxes, points, and oriented ellipses. It also introduces an automated annotation feature using the DEXTR algorithm, enhancing the labeling process significantly. With its user-friendly features and robust compatibility, Django Labeller makes image labeling a streamlined experience.

Features

Multi-Framework Compatibility: Compatible with Django, Flask, and Qt, allowing for flexibility in application development.
Various Annotation Options: Supports polygon, box, point, and oriented ellipse annotations to cater to different labeling needs.
Advanced Editing Capabilities: Polygonal labels can be edited with painting and boolean operations, providing precise control over annotations.
DEXTR Algorithm Integration: This feature allows for automatic generation of polygonal outlines of objects, making labeling faster and more efficient with minimal user input.
Schema Editor (New in v0.3): A dedicated schema editor lets users easily modify label classes, ensuring adaptability to different projects.
Easy Local Setup: Users can run a Flask or Qt application on their local machine without hassle, simplifying the installation process.
Desktop Application Functionality: The Qt-based desktop app can select directories of images for labeling, making the tool accessible for offline use.
Background Flask Server: The Qt application utilizes a background Flask server to manage HTML and static file delivery, streamlining the user experience.

Britefury

Django Labeller