theo2021/OnDA

Source code for "Online Unsupervised Domain Adaptation for Semantic Segmentation in Ever-Changing Conditions", ECCV 2022. This is the code has been implemented to perform training and evaluation of UDA approaches in continuous scenarios. The library has been implemented in PyTorch 1.7.1. Some newer versions should work as well.

27
/ 100
Experimental

This project helps machine learning engineers working on semantic segmentation tasks to improve model performance when deployment environments change. It takes a pre-trained semantic segmentation model and new, unlabeled image data from an evolving environment (like a camera moving between sunny and rainy conditions), and continuously adapts the model. The output is a more robust segmentation model that maintains accuracy over time without needing human re-labeling of new data.

No commits in the last 6 months.

Use this if you need to maintain high accuracy for semantic segmentation models in real-world applications where visual conditions or domains are constantly shifting, such as in autonomous driving or robotics.

Not ideal if your segmentation task involves static environments with consistent visual data, or if you are not working with deep learning models in PyTorch.

semantic-segmentation computer-vision model-adaptation machine-learning-engineering robotics
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 7 / 25
Maturity 16 / 25
Community 4 / 25

How are scores calculated?

Stars

28

Forks

1

Language

Python

License

GPL-2.0

Last pushed

Jun 11, 2024

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/theo2021/OnDA"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.