One of our focuses beside computer vision is data association, which is necessary for example for SLAM to detect loops or more general to see which feature you have already seen.
To encourage people to further develop their data association algorithms we provide a fully documented interesting and challeging data set. People can test their algortihms and compare them to other using the same data basis.
The data set evolved from the Ph.D. Thesis from Udo Frese and was recorded at the DLR (Deutsches Zentrum für Luft- und Raumfahrt). All processing has been done in the A7-[FreePerspective] project of SFB-TR/8 Spatial Cognition.
Inside Data Association (Workshop)
We organized a workshop on that topic at the RSS 2008, Zürich. Which was very well attended and had a lot of very good presentations. Have a look at the web page:
The DLR-Spatial Cognition Data Set
For comparing and benchmarking different algorithms for data association we maintain a interesting and challenging data set.
It has its own page at DLR-Spatial Cognition Data Set