## Perception as Bayesian InferenceIn recent years, Bayesian probability theory has emerged not only as a powerful tool for building computational theories of vision, but also as a general paradigm for studying human visual perception. The Bayesian approach provides new and powerful metaphors for conceptualizing visual perception, suggests novel questions to ask about perceptual processing, and provides the means to formalize theories of perception that make testable predictions about human perceptual performance. This book provides an introduction to and critical analysis of the Bayesian paradigm. Chapters by leading researchers in computational theory and experimental visual science introduce new theoretical frameworks for building perceptual theories, discuss the implications of the Bayesian paradigm for psychophysical studies of human perception, and describe specific applications of the approach. The editors have created a critical dialogue of ideas through the authors' commentaries on each others' chapters, conveying to the reader a unique appreciation for the issues and ideas raised in the book. |

### What people are saying - Write a review

User Review - Flag as inappropriate

10/29/2007 Found in Reiter's.

### Contents

Introduction D C Knill D Kersten A Yuille | 12 |

A unifying perspective D Mumford | 25 |

Modal structure and reliable inference A Jepson W Richards | 63 |

Priors preferences and categorical percepts W Richards A Jepson | 93 |

Bayesian decision theory and psychophysics A L Yuille | 123 |

Observer theory Bayes theory and psychophysics B M Bennett | 164 |

Commentaries | 213 |

Implications of a Bayesian formulation of visual information | 239 |

Ideal observers and human psychophysics | 306 |

A computational theory for binocular stereopsis P N Belhumeur | 323 |

The generic viewpoint assumption in a Bayesian framework | 365 |

The perception of shading and reflectance E H Adelson | 409 |

Banishing the homunculus H Barlow | 425 |

Commentaries | 451 |

507 | |

### Other editions - View all

### Common terms and phrases

algorithms analysis assume ball Bayes Bayesian approach Bayesian framework Bayesian inference cast shadow chapter coding competence observer compression Computer Vision configuration consider constraints context contours corresponding defined depth derived disparity domain domain theories elongation estimate example Figure Fisher Information Geman geodesic given homunculus human visual system hypothesis ideal observer image data Jepson kernel Kersten Knill light source likelihood function loss function Marr measure modal modes noise orientation particular pattern theory perceived perception perceptual inference possible posterior distribution posterior probability prior assumptions prior distribution prior model prior probability probabilistic probability distribution problem provides psychophysical random reflectance function regularities representation result scene interpretations scene parameters scene probability equation scene properties sensory shape from shading signal smooth space specific statistical stereo stereogram stereopsis structure surface normal surface shape task texels Theorem transparent variables velocity viewpoint visual perception visual system Yuille zero