Description
- Abstract:
- Perception of the 3D world is intuitive and effortless. The visual system is capable of processing the complexities in the retinal image to produce percepts of relative depths and absolute distances so effectively that healthy individuals have no problem navigating in and interacting with the world. To accomplish this, the visual system interprets and integrates multiple visual cues to depth, taking advantage of the diverse information available from moment to moment on the retina. Here, we investigated whether this process of interpretation and integration of depth cues is accomplished through a deterministic, non-probabilisic process or a stochastic, probabilistic process. To accomplish this, we scrutinized the motivations and derivations of two contradicting models, the intrinsic constraint theory and the Bayesian inference framework, which respectively take on these two approaches. We then analyzed experiments designed to tease apart the predictions of these two models. These experiments tested observers’ perceptions of depth using two general methods. The first determined the magnitude of perceived depth through adjustment tasks where observers adjusted a 2D icon to report their estimated depth occupied by a virtually presented stimulus. These experiments allowed us to examine how including visual cues to depth changes the reported perceived depth. The second method tested the precision by which observers could discriminate which virtual object was rendered with more depth. In these tasks, observers view a fixed standard stimulus and perform discrimination judgments until a variable comparison stimulus is consistently chosen as being deeper. Similarly, this method allowed us to test how the number and quality of visual cues influence the precision of the visual system. We found two main results. First, introducing more cues to a virtual rendering of an object increases the amount of depth participants reported, indicating systematic biases in perceived depth. Second, discrimination precision reflected the magnitude of the perceived depth of the comparison stimulus and did not depend on the standard stimulus. We found that the intrinsic constraint model offered the more parsimonious explanation for the data gathered via these two paradigms, suggesting that the visual system produces deterministic depth responses via a non-probabilistic integration strategy.
- Notes:
- Thesis (Ph. D.)--Brown University, 2023
Citation
Kemp, Jovan,
"Mechanisms Underlying 3D Vision: A comparison between stochastic and deterministic accounts"
(2023).
Cognitive, Linguistic, and Psychological Sciences Theses and Dissertations.
Brown Digital Repository. Brown University Library.
https://repository.library.brown.edu/studio/item/bdr:8r8m4ze2/
Relations
Collection:
-
Cognitive, Linguistic, and Psychological Sciences Theses and Dissertations
Theses and Dissertations for the Cognitive, Linguistic, and Psychological Sciences department....