** Download **

This challenging dataset contains outdoor images of 19 classes of different animals. All the images of this dataset are manually segmented into foreground and background region. Below the number of images per each class, some sample images showing the variations of the dataset and sample manual annotation of the dataset can be seen. The related publication to this dataset is

Joint Visual Vocabulary for Animal Classification
Heydar Maboudi Afkham, Alireza Tavakoli Tar ghi, Jan-oluf Eklundh, and Andrzej Pronobis
In Proceedings of the International Conference on Pattern Recognition (ICPR08), Tampa, FL, USA,
December 2008. 
[ PDF ]