Preface

Artificial vision is a rapidly growing discipline, aiming to build computational models of the visual functionalities in humans, as well as machines that emulate them. Although the ultimate goals are ambitious, successful theories and applications have appeared so far in the literature, thanks to the converging contributions from image processing, computational geometry, optics, pattern recognition and basic computer science.

The on-going research work suggests that vision is to be studied in the context of other – seemingly distinct – functionalities. In particular, active vision focuses on the dynamics of sensing, so that moving, modifying sensor parameters, and acquiring information concur to perform visual tasks in a purposive way. Moreover, quite often visual modules are to be included into more complex ‘intelligent’ systems devoted to real-world applications – e.g., automated control, inspection and diagnosis. Models of the objects are to be acquired and stored; plausible hypotheses are to be maintained about the scene content. In this way, profound links emerge between visual and cognitive capabilities, like learning, reasoning and communicating. Visual communication, in itself, involves a number of challenging topics, with a dramatic impact on the contemporary culture, where human–computer interaction and human dialogue via computers play a more and more significant role.

The present volume contains a number of selected review articles concerning the research trends mentioned above. In particular, Part I groups contributions in active vision, Part II deals with the integration of visual with cognitive capabilities, while Part III concerns visual communication.

We hope that the topics reported in this volume will encourage further research work along the emerging directions, and towards an integrated, comprehensive study of vision and intelligence in humans and machines.

Part of the material has been presented at the fourth School on Machine Vision, organized by the Italian Chapter of the International Association for Pattern Recognition (IAPR), and held in Udine, Italy, 24–28 October 1994.

We wish to thank all the researchers who contributed to the success of the school and the preparation of this volume. In particular, we thank all the members of the scientific committee: L. Cordella (Napoli), S. Gaglio (Palermo), S. Impedovo (Bari), G. Pieroni (Udine) and R. Stefanelli (Politecnico di Milano). The International Centre for Mechanical Sciences (CISM) is gratefully acknowledged for hosting the school. We are also grateful to Forum (Udine) for the help in typesetting and preparing the manuscripts.