MODE�BASEDINFLUENCE DIAGRAMS FOR MACHINE VISION

We show the soundness of automated con trol of machine vision systems based on in cremental creation and evaluation of a par ticular family of influence diagrams that rep resent hypotheses of imagery interpretation and possible subsequent processing decisions. In our approach, model-based machine vi sion techniques are integrated with hierarchi cal Bayesian inference to provide a framework for representing and matching instances of ob jects and relationships in imagery, and for ac cruing probabilities to rank order con:liicting scene interpretations. We extend a result of Tatman and Shachter to show that the se quence of processing decisions derived from evaluating the diagrams at each stage is the same as the sequence that would have been derived by evaluating the final influence dia gram that contains all random variables cre ated during the run of the vision system. I