LabelImg
A desktop tool for drawing labels on images to build training datasets. Archived; no longer maintained.
- Stars
- 25k
- Forks
- 6.6k
- Licence
- MIT
- Language
- Python
- Last update
- 2.4 years ago
- Type
- Dev tool
This repository is archived and no longer maintained.
What it is
LabelImg is a graphical image annotation tool written in Python. You open a folder of images and draw boxes around objects, then tag each box with a name, producing label files that machine learning models can train on. The project is archived and its description points users to Label Studio, a newer open source labeling tool, as the successor.
- Draw bounding boxes around objects in a folder of photos to build an object detection dataset
- Assign class names to labeled regions for image classification training data
- Review and correct existing label files in a local image folder
Great for
- People who need a simple local tool to draw bounding boxes on a small image set
- Students following older computer vision tutorials that use LabelImg
- Anyone who wants a basic, offline annotation tool and accepts that it is no longer updated
Not ideal for
- Teams starting a new labeling project, because the project is archived and points to Label Studio instead
- Projects needing text, audio, video, or time-series labeling, since the description only mentions image annotation
- People who need active bug fixes or support for newer Python versions
How to get it
- 1
Clone and build from source
You'll need: Python 3
Clone the repository
git clone https://github.com/HumanSignal/labelImg.git - 2
Run it
- Open the labelImg folder you cloned
- Install the Python dependencies listed in the repository
- Launch the application from the project folder
- Open a folder of images and start drawing boxes
- 3
Get the most out of it
- Check the repository for the latest setup instructions before installing, since the README was truncated
- Consider Label Studio, the project's recommended successor, for new work
- Use the MIT license terms if you plan to modify the code
Downloads
Strengths
- Simple, focused interface for image labeling
- MIT license allows free use and modification
- Widely known, with a large existing user base and many tutorials
Trade-offs
- Archived and no longer actively developed, so bugs and compatibility issues will not be fixed
- The README provided is truncated, so setup details could not be confirmed here
- Image labeling only, according to the project description
How it compares
| Project | Stars | Type | Level | Updated | Best for |
|---|---|---|---|---|---|
| labelImg (this) | 25k | Dev tool | beginner | 2.4 years ago | People who need a simple local tool to draw bounding boxes on a small image set |
| cvat | 17k | App | intermediate | 4 days ago | Computer vision teams that need to label images or video for model training |
| albumentations | 15.3k | Library | beginner | 1.4 years ago | Machine learning engineers training image classification or segmentation models with limited data |
| sherlock | 95.4k | CLI | intermediate | 4 days ago | Security researchers doing authorized open-source intelligence work |
| Swin-Transformer | 16.1k | AI model / agent | advanced | 2.3 years ago | Researchers reproducing results from the Swin Transformer paper |
| pytorch-book | 12.8k | Learning resource | beginner | 2.9 years ago | Chinese-speaking beginners learning PyTorch from scratch |
Stats synced from GitHub. The assessment above was drafted with AI from the project's README on 2026-10-29 and may miss recent changes — check the README before relying on a command. MIT licence: free to use, modify and distribute, including commercially, as long as you keep the copyright notice.