Learning Object Recognition in a NeuroBotic System
Rebecca Fay, Ulrich Kaufmann, Friedhelm Schwenker, Günther Palm
Object localisation and identification is a crucial problem for
advanced mobile service robots. We developed an object recognition
system that localises and identifies objects using a colour-based
visual attention control algorithm and a hierarchical neural network
for object classification utilising hierarchical class grouping. The
approach is evaluated in a test scenario where a robot is situated in
front of a table. The robot has to identify and manipulate objects
lying on this table. We evaluated the total object recognition
performance and compared the effectiveness of different feature
sets. The approach showed very encouraging results and meets real-time
constraints.