作者: V. S. S. Hwang , T. Matsuyama , L. S. Davis , A. Rosenfeld
DOI: 10.21236/ADA139874
关键词:
摘要: Abstract : This document describes a control structure for building an Image Understanding System. system can deal with objects diverse appearances when consistent spatial relations exist between objects. By accumulating predictions originated from existing instances, our dynamically reason about what to do in order construct interpretations of the image. paper, discusses parts proposed - representation knowledge, accumulation evidence, focus attention mechanism, and integration constraints top-down control. (Author)