metadata

dataset_info:
  features:
    - name: image
      dtype: image
    - name: label
      dtype: image
    - name: color_mask
      dtype: image
  splits:
    - name: train
      num_bytes: 229712082
      num_examples: 555
  download_size: 229827876
  dataset_size: 229712082
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
license: cc-by-nc-4.0
task_categories:
  - image-segmentation

UEC-Ramen555

https://mm.cs.uec.ac.jp/UEC-Ramen555/

Dataset Info

UEC-Ramen555 is a ramen image dataset that has 555 ramen images. Each image has segmentation mask of ramen attributes.

The semantic mask consists of 11 classes of pixel-level semantic labels that are manually annotated. And 5 additional soup category label annotations are included.

The ramen image semantic label contains semantic labels for the background, bowl, soup, spoon, chopsticks and toppings (e.g. half-cut egg, seaweed, sliced roast pork, etc.). Soup category labels include salt soup, soy sauce soup, miso soup, pork bone soup and spicy soup.

Citation

@inproceedings{ramenasyoulike2019,
  title={Ramen as You Like: Sketch-based Food Image Generation and Editing},
  author={Cho, Jaehyeong and Shimoda, Wataru and Yanai, Keiji},
  booktitle={ACM International Conference on Multimedia},
  year={2019}
}