MoCA: Moving Camouflaged Animals dataset

141 Videos, 37K frames

The largest video dataset for camouflaged animals discovery

Highly Challenging

67 categories of animals which have mastered camouflage

Annotations

We provide both bounding box annotations and motion type labels

Motion labels

3 main types of motion:

Locomotion: when the animal engages in a movement that leads to a significant change of its location within the scene.
Deformation: when the animal engages in a more delicate movement that only leads to a change in its pose while remaining in the same location.
Still: when the animal remains still.

Download

MoCA dataset (~10GB)

Terms of Use

The MoCA dataset is available to download for commercial/research purposes under a Creative Commons Attribution 4.0 International License. The copyright remains with the original owners of the video. A complete version of the license can be found here.

Please contact the authors below if you have any queries regarding the dataset.

Privacy and Ethical Responsibility

The MoCA dataset contains sequences of camouflaged animals in their natural habitat collected from the YouTube platform. These segments are extracted from documentaries and other shared content of YouTube channels. If you are the owner of a YouTube channel of which the content was included in our dataset, and do not wish for your sequence to be included in our benchmark, please use this form to request the deletion.

Note that none of the sequence included in our benchmark contains occurrence of humans and thus it is compliant with the General Data Protection Regulation (EU GDPR).

Publications

Please cite the following if you make use of the dataset:

Hala Lamdouar, Charig Yang, Weidi Xie, Andrew Zisserman

Betrayed by Motion: Camouflaged Object Discovery via Motion Segmentation

ACCV, 2020

Bibtex | Abstract | PDF | arXiv

               @Article{Lamdouar20,
                author       = "Hala Lamdouar and Charig Yang and Weidi Xie and Andrew Zisserman",
                title        = "Betrayed by Motion: Camouflaged Object Discovery via Motion Segmentation",
                journal      = "Asian Conference on Computer Vision",
                year         = "2020",
                }