We present a jhmework for determining probability distributions ower the space of possible image feature pupings. Such a jhmework allows higher level processes to m o n over manly plawible perceptual p u pings in an image, mther than committing to a specific image segmentation in the eadly stages of processing. We first derive an eqression for the probability that a set of features should be grouped together, conditioned on the observed image data associated with those fecrtums. This

more »

... measure formalizes the principle that featuw in an image should be grouped together when they participate in a common underlying geometric structure. We then present a representation scheme in which only those groupings with high probability are eqlicitly represented, while lafge sets of unlikely grouping hypotheses are implicitly represented. We present @mental w u l t s for a variety of real intensity images.