https://github.com/guy1998/synthetic-data-generator-for-event-cameras
Automated workflow for the generation of synthetic data that can be used as an input for event cameras based on previously taken videos
https://github.com/guy1998/synthetic-data-generator-for-event-cameras
computer-vision event-cameras numpy opencv synthetic-data
Last synced: 2 months ago
JSON representation
Automated workflow for the generation of synthetic data that can be used as an input for event cameras based on previously taken videos
- Host: GitHub
- URL: https://github.com/guy1998/synthetic-data-generator-for-event-cameras
- Owner: guy1998
- License: mit
- Created: 2025-03-19T20:13:45.000Z (2 months ago)
- Default Branch: main
- Last Pushed: 2025-03-19T21:05:27.000Z (2 months ago)
- Last Synced: 2025-03-19T21:33:48.021Z (2 months ago)
- Topics: computer-vision, event-cameras, numpy, opencv, synthetic-data
- Language: Python
- Homepage:
- Size: 11.7 KB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Synthetic data generator for event cameras
I am providing this workflow for those that need a light weight option for generating synthetic data as opposed to traditional options such as V2E or ESIM. These platforms are great and I invite you to have a look at them. However, if you are like me and cannot get your hands on a very advanced setup that can support such high end platforms you could use this code snippet.
The code is pretty easy to understand. We can run the file event_camera_sim.py to generate the coordinates, timestamps and polarities of the events (movements) happing in a certain video which we can provide as an argument. The results are stored in a .txt file the path of which we can also specify as an argument. To run this file you can use the following command:
```
python event_camera_sim.py input_video.avi --threshold 30 --output events.txt
```
Of course the output path can be chosen at your will. You can either use the .txt file which already contains the coordinate. Or you can run the coordinates-to-video.py to generate a synthetic data video. Make sure to update the path of the .txt on the line 5 of this file. To illustrate I am providing an example below using the video that features in the V2E tutorial.
Do not forget to run check the requirements.txt for the libraries you need to have. Hope this code will help you with your work! Happy innovation!