https://github.com/mearman/ml-frabeller
A Python-based video frame labelling tool aimed at labelling video datasets for machine learning model training. This tool uses OpenCV for its GUI and to allow frame by frame playback.
https://github.com/mearman/ml-frabeller
Last synced: 2 months ago
JSON representation
A Python-based video frame labelling tool aimed at labelling video datasets for machine learning model training. This tool uses OpenCV for its GUI and to allow frame by frame playback.
- Host: GitHub
- URL: https://github.com/mearman/ml-frabeller
- Owner: Mearman
- License: gpl-3.0
- Created: 2020-07-10T09:05:01.000Z (almost 5 years ago)
- Default Branch: master
- Last Pushed: 2023-02-23T15:01:02.000Z (over 2 years ago)
- Last Synced: 2025-04-04T16:51:22.128Z (3 months ago)
- Language: Python
- Homepage:
- Size: 57.6 KB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 6
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Introduction
ML-Frabeller is a Python based video frame labelling tool aimed at labelling video datasets for machine learning model training.
This tool uses OpenCV for its GUI and to allow frame by frame playback.
This project is based on [maximus009](https://github.com/maximus009)'s Python OpenCV [video player](https://github.com/maximus009/VideoPlayer) that was written in 2015 and updated in 2019 [here](https://github.com/839687571/VideoPlayer).
The video can be played through with variable FPS and skipped forward and backward frame-by-frame.
When scrubbing through frames, arbitrary labels of 0 - 9 can be assigned with the 0-9 keys on a frame-by-frame basis. The labels are automatically exported and load to and from a csv in the same directory as the video.
## Dependency installation
The dependencies can be install in several ways.
### Using [setup.py](./setup.py)
```
pip install .
```
### Using [requirements.txt](./requirements.txt)
```
pip install -r requirements.txt
```
### Manually
```
pip install opencv-python
pip install pandas
pip install numpy
```
## Usage:
`$ python frame_labeller.py videos/video_file.mp4`
When no params are passed, you'll be prompted to provide a valid file
```0
$ python frame_labeller.py
No file name argument provided
Full video path:
```
## Keyboard commands
```
Space Play/Pause
Left/Right Jump 1 frame
Up/Down Jump 5 frames
+/- Change FPS
c Capture frame
s Save csv of labels
0-9 Toggle label 0-9
esc/q Quit
```