Datasets:

nobg
/

camo

dataset_info:
  features:
    - name: image_name
      dtype: string
    - name: image
      dtype: image
    - name: mask
      dtype: image
    - name: overlaid_mask_1
      dtype: image
    - name: overlaid_mask_2
      dtype: image
  splits:
    - name: train
      num_bytes: 236064088
      num_examples: 1000
    - name: test
      num_bytes: 68819435
      num_examples: 250
  download_size: 287318302
  dataset_size: 304883523
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: test
        path: data/test-*
license: cc-by-nc-4.0
pretty_name: CAMO
task_categories:
  - image-segmentation
tags:
  - camouflaged-object-detection
  - camouflage
  - segmentation
  - background-removal
size_categories:
  - 1K<n<10K

CAMO (Camouflaged Object dataset)

Camouflaged object segmentation dataset, mirrored to the nobg org for convenience. Each example is an RGB image of a camouflaged object paired with its binary ground-truth mask.

image_name: original file name
image: RGB source image
mask: binary camouflaged-object ground-truth mask
overlaid_mask_1, overlaid_mask_2: visualizations of the mask overlaid on the image

Splits: train (CAMO-TR) and test (CAMO-TE).

Source & credit

From the CAMO dataset introduced in "Anabranch Network for Camouflaged Object Segmentation" (Le et al., CVIU 2019). Project page: https://sites.google.com/view/ltnghia/research/camo. Redistributed here for research / non-commercial use.

Citation

@article{le2019anabranch,
  title={Anabranch Network for Camouflaged Object Segmentation},
  author={Le, Trung-Nghia and Nguyen, Tam V. and Nie, Zhongliang and Tran, Minh-Triet and Sugimoto, Akihiro},
  journal={Computer Vision and Image Understanding},
  volume={184},
  pages={45--56},
  year={2019},
  publisher={Elsevier}
}