Software agents for computer vision: a preliminary discussion

Abstract

We start a preliminary discussion about the development of software agents developed for computer vision. Such agents must be heterogeneous and adaptive in nature to satisfy increasing requests for integration-cooperation capability exhibited by current computer vision applications (e.g. biomedical imaging and image retrieval on the Web). We give a brief introduction to the overall architecture design and we focus on lowest level agents, called W-bots (Working Software Robots). We define the nature of such agents and describe their ways of interaction, and their adaptive capabilities. Finally, we present an example of the proposed architecture utilizing VDM++ to specify agents as active objects.