SceneBank Feature: You can also concatenate all the numbers return for scene classification and use this as a feature vector of an image for other tasks (e.g. attribute recognition). We call this feature SceneBank (as a counterpart to ObjectBank by Li et al.)
Goal: The goal of this website and API is to maintain a free service to demonstrate the state-of-the-art performance for visual scene understanding in computer vision. We aim to keep this system updated with the best algorithm that we can implement with reasonable speed and is trained using the latest available SUN Database. Therefore, the performance will improve from time to time when we try to train a better model.
Please cite the following paper if you use this service.
J. Xiao, J. Hays, K. Ehinger, A. Oliva, and A. Torralba.
SUN Database: Large-scale Scene Recognition from Abbey to Zoo. IEEE Conference on Computer Vision and Pattern Recognition (CVPR2010)