qizhuli/Weakly-Supervised-Panoptic-Segmentation

Weakly- and Semi-Supervised Panoptic Segmentation (ECCV18)

42
/ 100
Emerging

This project helps computer vision researchers and practitioners efficiently generate high-quality pixel-level annotations for urban street scenes. It takes images with basic scene-level tags and bounding boxes as input. It then outputs detailed 'panoptic' segmentation masks, which classify every pixel and differentiate between individual objects. This tool is for those who need precise image segmentation without the cost of manual pixel-level labeling.

162 stars. No commits in the last 6 months.

Use this if you need to create detailed pixel-level (panoptic) segmentation masks for urban images, but only have access to cheaper, less precise labels like image-level tags and bounding boxes.

Not ideal if you already have fully-annotated datasets or are working with domains outside of urban street scenes, as it's specifically tailored for Cityscapes-like data.

urban-scene-understanding image-annotation autonomous-driving computer-vision-research semantic-segmentation
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 10 / 25
Maturity 16 / 25
Community 16 / 25

How are scores calculated?

Stars

162

Forks

22

Language

MATLAB

License

MIT

Last pushed

Jun 08, 2021

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/computer-vision/qizhuli/Weakly-Supervised-Panoptic-Segmentation"

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