ICARO

Image Collection of Annotated Real-world Objects

Overview

ICARO (Image Collection of Annotated Real-world Objects) is a dataset of images especially designed for category-level object recognition, detection and pose estimation. The goal of ICARO is to establish a benchmark for investigating the performance of different methods, over a wide spectrum of real-world images.

Database Rights

The database has been made publicly available for scientific research purposes.

ICARO data includes images downloaded from the flickr website. Use of these images must respect flickr terms of use.

Examples

26 classes
apple banana bike bottle camera car chair dustbin foot
apple banana bike bottle camera car chair dustbin foot
apple banana bike bottle camera car chair dustbin foot
apple banana bike bottle camera car chair dustbin foot

fork glasses hand hat knife laptop mobile monitor motorbike
fork glasses hand hat knife laptop mobile monitor motorbike
fork glasses hand hat knife laptop mobile monitor motorbike
fork glasses hand hat knife laptop mobile monitor motorbike

mug phone pistol potted plant shoe sofa tel. cabin traffic lights
mug phone pistol potted plant shoe sofa telephone cabin traffic light
mug phone pistol potted plant shoe sofa telephone cabin traffic light
mug phone pistol potted plant shoe sofa telephone cabin traffic light

Download

ICARO 2010

  • Download the development kit (~190 MB). It contains: training and test data, and the software.Last Update: 31-March-11

How to Cite ICARO?

We would appreciate it if you cite our work when using the ICARO dataset

@MISC{icaro2010,
  author =	 {Lopez-Sastre, R.~J. and Redondo-Cabrera, C. and Gil-Jimenez, P. and Maldonado-Bascon, S.},
  title =	 {{ICARO}: {I}mage {C}ollection of {A}nnotated {R}eal-world {O}bjects},
  howpublished = {\url{http://agamenon.tsc.uah.es/Personales/rlopez/data/icaro}},
  year = {2010}
}

Acknowledgements

The authors gratefully acknowledge support by the Collaboration Grants 2008-2009, 2009-2010 and 2010-2011 of the "Ministerio de Educación" of Spain. Here we go with the list of people who spent many long hours annotating images: Ana Zaldívar, Carolina Redondo, Jorge Tarlea, Josu García, Vanesa Martín, Antonio Vegue and Isabel Ventura. Thanks! ;)

Contact

Please, direct all your comments, questions, bugs at Roberto J. López-Sastre.