Advances in Imaging and Electron Physics, Volume 117

The four articles that make up this latest addition to this series deal with very new developments in signal and image processing and recent thinking about the notion of resolution. We begin with a long-standing concern of image analysis, namely, granulometry. Granulometric Titering was one of the earliest preoccupations of the founders of mathematical morphology, a subject to which E. R. Dougherty has contributed extensively. Here, he and Y.-d. Chen describe at length a procedure for the optimal and adaptive design of logical granulometric Titers, with examples from character-recognition and blood-cell analysis.

Wavelets have attracted enormous attention since they entered the signal processing toolbox some years ago. G. Evangelista, to whom we owe many important developments on this subject, here presents this class of wavelets in detail, thereby making these new ideas accessible to a wide audience. This long chapter forms a complete monograph on the subject.

The median fiiter was 1st used empirically in image processing in an attempt to smooth noisy inrages without undue blurring of sharp contrast changes (edges in particular). G. R. Arce has been very active in elucidating the underlying structure of these fiiters, and here he and J. L. Paredes describe in detail the principles of this family of fiiters and. in particular, recent developments in fiitering and smoothing based on threshold decomposition.

The final chapter is a major contribution to our understanding of the notion of resolution. A. van den Bos and A. J. van Dekker have examined these difficulties with great care, and an alternative approach to resolution has emerged from their work, based on parametric models of the observation process and maximum likelihood methods of parameter estimation.