chen742/DCF

This is the official implementation of "Transferring to Real-World Layouts: A Depth-aware Framework for Scene Adaptation" (Accepted at ACM MM 2024 as Oral).

27
/ 100
Experimental

This project helps computer vision researchers and engineers improve the accuracy of semantic scene segmentation models, especially when high-quality real-world labeled data is scarce. It takes synthetic images (like from games or simulations) and unlabeled real-world images (like street scenes) to produce a more precise scene segmentation model capable of identifying objects like buildings, sidewalks, and sky in real photographs.

No commits in the last 6 months.

Use this if you need to train a robust scene segmentation model for real-world applications but only have access to synthetic data with labels and a large amount of unlabeled real-world images.

Not ideal if you already have perfectly labeled real-world datasets for your target domain or if your task doesn't involve adapting models trained on synthetic environments.

computer-vision semantic-segmentation unsupervised-domain-adaptation synthetic-data image-analysis
No License Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 5 / 25
Maturity 8 / 25
Community 14 / 25

How are scores calculated?

Stars

14

Forks

3

Language

Python

License

Last pushed

Aug 24, 2024

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/diffusion/chen742/DCF"

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