Representing, perceiving and remembering the shape of visual space.

Oliva, Park, & Konkle

Humans constantly engage in automatic and rapid analysis of spatial scene structure when navigating an environment or searching for objects. This chapter draws on current research from computational, behavioral and neuroscience perspectives aimed at understanding how the human brain perceives, represents, and remembers the shape of space. Space can be described with both structural descriptions, which reflect layout of surfaces in the physical world, and semantic descriptions, which incorporate an observers understanding of the environment. We first review two formal approaches which can quantify the structural properties of scene layout: the isovist representation and the spatial envelope representation. Next we explore how space is experienced by observers, by reviewing behavioral results in which different factors distort the perceived space away from a veridical representation. Finally, we examine how representations of specific views of space are maintained in memory, and discuss potential neural mechanisms involved in integrating views into larger environments. Gaining a formal understanding of how geometric aspects of space map onto human perceptual, cognitive, and neural systems will help to create efficient functional spaces for our daily interactions with the world.




Oliva, A., Park, S., & Konkle, T. (2010). Representing, Perceiving and Remembering the Shape of Visual Space. Computational Vision in Neural and Machine Systems. Cambridge University Press, edited by Laurence R Harris and Michael Jenkin.