A considerable amount of work has been done on the topic of aspect graphs in the
past two decades. Unfortunately, the different approaches have remained quite
theoretical and of little use for object recognition problems in the real
world. A novel approach is presented in this report whereby the aspect graph of
an object is constructed based on its extrinsic properties with respect to a
database of objects. The approach is inspired by recent work on active object
recognition by Arbel and Ferrie [
1,
3,
2]
and the resulting aspect graph formulation is in turn used to define
a new gaze planning strategy. Experimental results indicate that the
performance of the new strategy is comparable with that of the strategy
developed by Arbel and Ferrie.
