Liang K-H, Tjahjadi T & Yang Y-H

Proc. UK Symp. on Applications of Time-Frequency and Time-Scale Methods

IEEE Signal Processing Chapter (UKRI Section), 1995, 58-65

Abstract

Edge detection attempts to reconstruct 3-D physical edges from a 2-D image. It is therefore an ill-posed problem, and a regularization procedure is required to convert it into a well-posed problem. This procedure introduces a regularization parameter alpha for adjusting the extent of the regularization effect: the larger the value of alpha is, the stronger the regularization effect. Therefore, alpha could be interpreted as a scale of the regularization effect in an image. In this paper, an alpha scale space is used to provide the basis for a multiscale edge detection scheme which employs the regularized cubic B-spline fitting.