Skip to content

Dataset · naufalso

Cityscapes with weather and seasonal effects for segmentation

naufalso/cityscape-adverse

A benchmark based on Cityscapes with 8 weather and lighting conditions while preserving original annotations.

What's inside

The dataset extends the well-known Cityscapes dataset using diffusion image editing, adding eight realistic conditions: rain, fog, spring, autumn, snow, sunny weather, night, and dawn. All generated scenes retain the original semantic segmentation masks at a resolution of 2048x1024.

What you can do with it

  • Testing the robustness of semantic segmentation models in adverse weather conditions
  • Fine-tuning perception systems for autonomous driving
Downloads this month
71.5K
Likes
3

How to load it

from datasets import load_dataset

ds = load_dataset("naufalso/cityscape-adverse")