Skip to content
Free planNo credit card requiredStart writing

report · 1993

Intermediate Levels of Visual Processing

Ken Nakayama

Cited by 0Abstract and details only

Abstract

Abstract : We completed a comprehensive theoretical account of visual surface representation based on a new understanding of many of our recently reported perceptual phenomena. In particular, we determined the degree to which the 'generic view principle' can explain these new findings and also extended its domain to a large number of perceptual phenomena (Shimojo and Nakayama, 1992). We apply the generic sampling principle which indicates that the visual system acts as if it were viewing surface layouts from generic, not accidental vantage points. Through the observer's experience of optical sampling, which can be characterize geometrically, the visual system makes associative connections between images and surfaces, passively internalizing the conditional probabilities of image sampling from surfaces. In turn, this enables the visual system to determine which surface a given image most strongly predicts. Thus, visual surface perception can be considered as inverse ecological optics based on learning through ecological optics. Because of its neglect of prior probabilities and higher level knowledge, visual surface perception deviates from strict Bayesian notions of inference.

Cite this paper

APA 7

Nakayama, K. (1993). Intermediate Levels of Visual Processing. https://doi.org/10.21236/ada284922

MLA 9

Nakayama, Ken. “Intermediate Levels of Visual Processing.” 1993, https://doi.org/10.21236/ada284922.

Chicago

Nakayama, Ken. 1993. “Intermediate Levels of Visual Processing.” https://doi.org/10.21236/ada284922.

Harvard

Nakayama, K. (1993) ‘Intermediate Levels of Visual Processing’. Available at: https://doi.org/10.21236/ada284922.

BibTeX
@article{nakayama1993intermediate,
  title = {{Intermediate Levels of Visual Processing}},
  author = {Nakayama, Ken},
  year = {1993},
  doi = {10.21236/ada284922}
}

Made from the index record. Check names, capitalisation and page numbers against the paper. Need another style? AI for Research has over 10,000.