Bark-101 is a challenging dataset made of 101 bark classes. It has been designed on top of the LifeCLEF challenge to evaluate both texture and bark recognition algorithms in the wild. Bark-101 answers the lack of extended and segmented dataset in the context of bark recognition.
If you find Bark-101 useful in your research, please, cite us :
R. Ratajczak, S. Bertrand, C. Crispim-Junior and L. Tougne, Efficient Bark Recognition in the Wild, International Conference on Computer Vision Theory and Applications (VISAPP'19), pp. 240-248, 25-27 February, Prague, Czech Republic, 2019. DOI: 10.5220/0007361902400248