Hugging Face's logo

Datasets:
isp-uv-es
/
IPL-Cityscapes-LuminanceContrasts

Tasks:
Image Segmentation
Modalities:
Image
Size:
n<1K
Tags:
image
image-segmentation
License:
Dataset card Viewer Files Community
IPL-Cityscapes-LuminanceContrasts / README.md
hercapa's picture
hercapa
Update README.md
5e1e4e4 verified
preview code
|
raw
history blame
1.01 kB
metadata
license: apache-2.0
task_categories:
  - image-segmentation
pretty_name: IPL-CityscapesLuminanceContrasts-dataset
size_categories:
  - n<1K
tags:
  - image
  - image-segmentation

IPL-CityscapesLuminanceContrasts-dataset

Controled luminance and contrasts modified dataset version of the famous autonomous driving semantic segmentation Cityscapes dataset.

Dataset generation

For each original image, we convert it to ATD color space. Once in this space, we compute its mean luminance, achromatic contrast and chromatic contast. We modify each of its characteristics in turns from 0.5 to 1.5 of its original value. Then we return the image to RGB space. We apply it to 100 Cityscapes images.

Dataset structure

The original_images folder contains the original 100 Cityscapes images and the labels folder contains the corresponding ground truth segmentation masks. The modified images are in the modified_images folder. This folder has 11 subfolder according to the modified chromatic contrast.