Guided sampling for rapid object location using biologically motivated model

Abstract

A known problem relating to the rapid location of objects by sampling is tackled. Specifically, a new guided sampling technique has been shown to be able to take into account a priori probabilities for object placement and to automatically carry out body-based or edge-based sampling as appropriate. In addition, the technique carries out full binary search tree edge location without explicit programming. Overall, the sequential attention-based analysis emulates certain desirable aspects of the operation of the human visual system.