Omkar M Parkhi and Andrea Vedaldi and Andrew Zisserman and C. V. Jawahar
Overview
We have created a 37 category pet dataset with roughly 200 images for each class. The images have a large variations in scale, pose and lighting. All images have an associated ground truth annotation of breed, head ROI, and pixel level trimap segmentation.
Downloads
The Oxford-IIIT Pet dataset and annotations are roughly 800 MB in size and available for download via BitTorrent with Academic Torrents:
We recommend the use of BitTorrent protocol. If its use is not possible, the dataset and annotations are also available for download over http as two separate files: images.tar.gz (dataset) and annotations.tar.gz (groundtruth data).
Annotations Examples
The following annotations are available for every image in the dataset: (a) species and breed name; (b) a tight bounding box (ROI) around the head of the animal; and (c) a pixel level foreground-background segmentation (Trimap).
Dataset Statistics
License
The dataset is available to download for commercial/research purposes under a Creative Commons Attribution-ShareAlike 4.0 International License. The copyright remains with the original owners of the images.
Relevant Publications
Acknowledgements
This work is funded by The UK India Education and Research Initiative (UKIERI) and ERC Grant VisRec.