The scale-space primal sketch
Scale-space theory provides a well-founded framework for dealing with image structures, which naturally occur at different scales. According to this theory, from a given signal one can generate a family of derived signals by successively removing structures when moving from finer to coarser scales. At any scale in this representation, features can then be defined either by geometric constructions or by combining Gaussian derivatives into differential invariants. In contrast to other multi-scale or multi-resolution representations, scale-space is based on a precise mathematical definition of causality, or scale invariance, and the behaviour of structure as scale changes can be analytically described. However, the information in the scale-space embedding is only implicit in the grey-level values. The smoothed images in the raw scale-space representation contain no explicit information about the features in them or the relations between features at different levels of scale.
KeywordsSaddle Point Scale Parameter Local Extremum Coarse Scale Discrete Signal
Unable to display preview. Download preview PDF.