Datasets‎ > ‎

Caltech Game Covers Dataset


This is a set of CD/DVD covers of video games. The set was downloaded from during the summer of 2008. The set includes around 11,400 images for games on different consoles, e.g. PC, Playstation, Xbox, Wii, ... etc.
The set has a lot of images that are similar to each other, which we call "image families". These image families are mainly the covers of the same game on different consoles, different languages, or different editions.
In order for this dataset to be useful, we sifted through the whole set and manually labeled those image families.
The archive includes the images and the label for every image. It also includes labels for a number of subsets with a certain number of minimum family size.



  • Mohamed Aly, Peter Welinder, Mario Munich, and Pietro Perona. Towards Automated Large Scale Discovery of Image Families, Second IEEE Workshop on Internet Vision, CVPR, Miami, Florida, June 2009. [pdf]
  • CaltechCV Family Viewer: to view/edit the families in the dataset.
  • Automatic Discovery of Image Families: more information about the project.