Special section on frontiers in machine vision applications introduction

This special secton consists of four revised papers that were originally presented at the IAPR-MVA' 92 (International Associaton for Pattern Recognition's Workshop on Machine Vision Applications) held in Tokyo, Japan, December 7-9, 1992. It was the third workshop in a series of biennial international workshops held in Japan that focus on machine vision applications. They are sponsored by TC-6 (Special-purpose architecture), TC-8 (Applications in industry) and TC-10 (Graphics recognition) of IAPR and supported by several Japanese academic societies including IEICE (Institute of Electronics, Information and Communcations Engineers). For MVA'92, 240 abstracts were submitted and 40 papers for oral presentation and 109 papers for poster presentation were accepted from both industry and schools. (The next workshop, IAPR-MVA' 94, is planned to be held in Kawasaki, Japan, December 13-15, 1994.) Dr. Masakazu Ejiri, editor in chief of Machine Vision and Applications, suggested that a special section be published in the MV&A journal composed of highquality papers selected from the IAPR-MVA'92 presentations. Thus, the editorial committee for this special section consisted of both MV&A editorial board members and IAPR-MVA'92 program committee members. As the IAPR-MVA'92 program chair, I have the honor to be the guest editor of this special section. Of the high-quality papers presented at the IAPR-MVA'92 workshop, the editorial committee dediced to recommend four papers for publication to the editorial board of Machine Vision and Applications. Meanwhile, nine other papers have been included in the October 1993 issue of IEICE transactions E76D as a special section on Machine Vision Applications. The four papers included in this special section were selected as delineating the frontiers of machine vision applications. They are concerned with a real-time vision system using a prototype integrated memory array processor LSI, a multiscale regularity measure as a geometric criterion for image segmentation, a robust algorithm for layout extraction of mixed mode documents, and motion estimation methods for arbitrary-shaped objects like clouds. I hope that this special section is informative to the readers of MV&A and that it contributes to the further development of machine vision applications.