The Object Recognition software is from Thomas Mauer. It was part of  his Master Thesis and he published his work under the terms of the GPL. It is developed in c++ and I only enhanced it to deliver the positions to the DeviceServer. The patch I did is rather raw hack and a basic proof of concept, and I hope that somebody out there will take care on this topic and make it production ready. It's not a hard job with good c++ knowledge, so how about you to voluteer ? More on the object recognition as usual in the OpenMSR User Guide or the master thesis from Thomas Maurer. Let's thank him for his work.