The simple act of seeing an object conceals an array of fascinating and difficult scientific questions. Since objects occlude themselves and each other, the information available for seeing objects is fragmentary. Yet we perceive whole objects, and doing so provides a key foundation for action, thought, language and learning. As we move, the visible parts of objects change, yet we experience these objects as being stable over time. The phenomenal coherence and stability of objects are simple facts of perceptual experience, but they result from complex computational processes. Despite years of work, researchers in biological and computational vision have yet to achieve satisfactory accounts of unit formation. The present period, however, is one of new insights and rapid progress. This book addresses the problem of how the human visual system organizes inputs that are fragmented in space and time into coherent, stable perceptual units - objects. In doing so it addresses the following questions: what kinds of segmentation and grouping abilities exist in human perceivers? What information and computational processes achieve segmentation and grouping?
What are the psychological consequences of perceiving whole objects? In an effort to give a comprehensive, integrative answer to these questions the volume includes chapters from authors in five areas: philosophical foundations; computational and neural models of segmentation and grouping; attention and grouping; development; segmentation and grouping over time (event perception). The chapters review the contemporary state of the field, describe recent theoretical and empirical research, and provide some explicit suggestions for future research direction. "From Fragments to Objects: Segmentation and Grouping in Vision" takes a comprehensive cognitive science approach to object perception, brings together separate lines of research in object perception in one volume, gives an integrated and up-to-date review of theory and empirical research and offers directions for future study.
Publisher: Elsevier Science & Technology
Number of pages: 622
Weight: 1100 g
Dimensions: 225 x 149 x 43 mm