NTU-Tree 51

Description

NTU-Tree 51 is a fine-grained image dataset collected for recognizing different tree species at a distance. All the images were cropped from Google Street View images, which were captured continuously from a distance on a moving vehicle. This dataset contains 2613 street view tree images in total, which is composed of 51 common tree species in Singapore, and each species contains 30-70 samples. This dataset is challenging because of its large intra-class variance, and relatively small inter-class variance.

Related Papers

• Zhen Zuo and Gang Wang. “Learning Discriminative Hierarchical Features for Object Recognition.” Signal Processing Letters (SPL), IEEE, 2014. [Link]
• Zhen Zuo and Gang Wang. “Recognizing trees at a distance with discriminative deep feature learning.” In Proceedings of the Information, Communications and Signal Processing (ICICS) 2013 9th International Conference on, pp. 1-5. IEEE, 2013. [Link]