A dataset of multiclass semantic segmentation image annotations for the first 250 images of the "Duckietown Object Detection Dataset".
This dataset defines 8 semantic classes (7 distinct classes + implicit background class):
Ego Lane
Ego Lane
The lane the agent is supposed to be driving in (default right-hand… See the full description on the dataset page:
https://huggingface.co/datasets/hamnaanaa/Duckietown-Multiclass-Semantic-Segmentation-Dataset.