Collections of images of everyday objects obtained with the iCub robot as it observes its environment, to support the evaluation of object recognition models. Learn about current research by the iCub team and access the iCub project site to download the image datasets. The dataset includes images of 200 objects divided into 20 categories, with 10 images per object. Each object is acquired while undergoing geometric transformations, in order to study invariance to changes in viewpoint.
The iCub project is the result of a collaboration between the Italian Institute of Technology, the Statistical Learning & Image Processing Group at Genoa University, and the Laboratory for Computational and Statistical Learning at MIT.
URL: https://robotology.github.io/iCubWorld/
Videos:
- Giorgio Metta, Introduction to the iCub robot (34:04)
- iCub Team (Carlo Ciliberto, Guilia Pasquale, Rafaello Camoriano, Alessandro Roncone), Overview of research on the iCub robot (1:05:35)
Additional Resources:
- Fanello, S. R., Ciliberto, C., Santoro, M., Natale, L., Metta, G., Rosasco, L. & Odone, F. (2013) iCub world: Friendly robots help building good vision data-sets, Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR).