An Algorithm for Matching of Distorted Waveforms Using a Scale-Based Description.
Hiroshi Sakou, Haruo Yoda, Masakazu Ejiri · 1988
ARSTRACT This paper proposes a matching algorithm for two In the process of visually matching two mutually-distorted waveforms, each having partial mutuallyaistorted waveforms, we first try to detect major differences in the hierarchical structure of its scale-space. corresponding points between them by observing their 'I'his algorithm is applied to matching of two-dimensional roc,gh outlines, Then, we look for,nore detailed shapes and to pattern-width measurement of semiconductor corresponding points from every division of the waveforms chip patterns obtained from a scanning electron microscope between the major points. This suggests that the human (SEM). observer usually matches corresponding portions of the waveforms by observing them a t different resolutionsor BASIC CONCEPT viewscales. The matching algorithm orowsed in this oaoer models this human matching p;ocbss for disioried waveforms, by using a scale-based description. The algorithm is composed of two main parts; one is the method of matching each characteristic point of two waveforms a t a certain scale factor, and the other is the methodof controlling the matching, by which each waveformis divided into smaller portions a t decreasing scale factors until the matching succeeds. The algorithm is also applied to matching of two-dimensional distorted shapes and to pattern-width measurement of semiconductor chip patterns obtained from a scanning electron microscope. The generality and practicability of this algorithm are confirmed through these experiments.