University of Minnesota
Visual Inference of Objects and Size: Contextual Illusions and Figure--Ground Decisions in Human Vision
Abstract
dc:description.abstractVision must recover stable object properties from image data that is noisy and underdetermined. This dissertation uses perceived size, contextual illusions, and figure-ground decisions as linked tests of how the visual system manages that problem. The central proposal is that size constancy, contextual distortion, and loss of figure status arise from coordinated stages of processing rather than from unrelated mechanisms. The first empirical chapter examines size constancy in a three-dimensional corridor with objects that differ in morphology. Corridor cues improve constancy for all shapes, but most strongly for objects whose geometry aligns with the corridor frame, suggesting that depth-based scaling depends on how well object structure fits the inferred spatial layout. The second empirical chapter asks how surround-based and depth-based influences on apparent size combine. Using factorial manipulations of Ebbinghaus context, carrier spatial frequency, and Ponzo-style perspective, it shows that Ebbinghaus magnitudes remain stable after baseline correction, whereas perspective cues shift overall perceived size without altering the surround effect-supporting a hierarchically dissociated two-stage architecture. The third empirical chapter studies figure-ground decisions in a Connect-Five occlusion paradigm. By varying support for a fragmented figure against competing background structure, it identifies conditions under which observers adopt a background-centered description, operationalizing perceptual surrender as a measurable tipping point under clutter. These results motivate a layered view in which early surround normalization, intermediate figure-ground gating, and late size-scaling jointly shape perceptual experience.
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Priewe, Jett
Rights
- Language dc:language.iso
- en
Identifiers
dc:identifier.*- Dc Identifier Citation
- Priewe, Jett. (2026). Visual Inference of Objects and Size: Contextual Illusions and Figure--Ground Decisions in Human Vision. Retrieved from the University Digital Conservancy, https://hdl.handle.net/11299/283140.