We present a paradigm for feedback strategies that find instances of a gene
ric class of objects bf improving on established single-pass hypothesis gen
eration and verification approaches. We improve upon the mechanisms of the
traditional or classical image processing systems by introducing control St
rategies at low, intermediate, and high levels of analysis. We produce opti
mal sets of low-level features to reduce the number of hypotheses generated
. The feedback further enables updated sets of features to be extracted so
that the target object may be located even in very noisy data. The use of a
n interest operator in the feedback directs the search through the hypothes
es in an optimal manner, so minimizing the amount of feedback to false alar
ms. Furthermore, we aim to obtain detailed information about a complex obje
ct and not just its location. Thus, following top-down recognition of the o
bject our feedback control directs the search for missing information. The
system can extract complex objects in a scale and rotation independent mann
er where the objects may be partially occluded. The method is illustrated u
sing box shaped objects and noisy IR images of a number of bridges.