🎯 AnnotationToolKit

A clean, open-source Python application for annotating object detection datasets - built entirely with PyQt5 - EASY TO USE.

---

✨ About

AnnotationToolKit is a lightweight, user-friendly annotation tool built entirely in Python with PyQt5. It's designed to help developers, researchers, and educators create high-quality object detection datasets — faster, easier, and fully offline.

Whether you're working with curated images or extracting frames from videos, AnnotationToolKit gives you full control over your annotation workflow — from labeling to export.

⚡ Most importantly! It runs entirely on your local machine, with no cloud dependencies or forced sign-ins - so your data stays private, secure, and yours.

---

🚀 Features

• 🎥 Video support – Load videos and extract frames on the fly
• 🖼️ Image annotation – Annotate bounding boxes across images and video frames
• 🧠 Custom class support – Create and manage your own class list dynamically
• 💾 Export to popular format – Compatible with common training pipelines
• 🧰 Pure Python – Easily modifiable and extendable
• 🪶 Lightweight & cross-platform – Runs anywhere Python runs

---

📦 Installation

✅ Requirements

• Python 3.7+
• PyQt5
• OpenCV

🔧 Setup

    git clone https://github.com/joelcic/AnnotationToolKit.git
    cd AnnotationToolKit
    python -m venv .venv
    source .venv/bin/activate   # Windows: .venv\Scripts\activate
    pip install -r requirements.txt
    python main.py

---

🧭 User Guide

AnnotationToolKit includes three main views: a welcome screen, an image/video annotation tool, and a video-to-frame extraction module. Here's how to use each part of the app:

🖼️ Annotation Tool

What it does:
Draw bounding boxes on images or video frames and export them in a popular dataset format.

How to use:

• Select your image folder using 🖼️ Add image source
• Choose an output folder via 📁 Add output source
• Add your classes using the text field + ➕ Add
• Draw boxes by clicking and dragging
• Choose class via dropdown popup
• Click 💾 Save to export
• Navigate with ◀ / ▶

🎥 Video Tool

What it does:
Extract and save specific frames from video files.

How to use:

• Click 🎥 Load Video to open a file
• Step through frames with ◀ / ▶
• Use ⏯ Play / Pause to toggle playback
• Set an output folder with 📁 Add output source
• Use 💾 Save Frame to export frames.

---

🧩 Upcoming Features

We're actively improving AnnotationToolKit to make it more powerful and flexible for all types of computer vision workflows. Here's what's on the roadmap:

• 🧠 Image classification mode

• ✏️ Polygon & segmentation tools

• 🤖 Automatic annotation/classification (AI-assisted labeling)

• 🗂️ Support for more export formats

• 🎨 Custom style preferences

• 🧱 Resizable & draggable bounding boxes

• 💼 Project saving/loading

Got an idea or need a feature? Open an issue — we’d love your input!

---

👤 Author

Made with ❤️ by Joel