Image-based object recognition in man, monkey and machine

Michael J. Tarr, Heinrich Bulthoff

Research output: Contribution to journalArticle

235 Citations (Scopus)

Abstract

Theories of visual object recognition must solve the problem of recognizing 3D objects given that perceivers only receive 2D patterns of light on their retinae. Recent findings from human psychophysics, neurophysiology and machine vision provide converging evidence for 'image-based' models in which objects are represented as collections of viewpoint-specific local features. This approach is contrasted with 'structural-description' models in which objects are represented as configurations of 3D volumes or parts. We then review recent behavioral results that address the biological plausibility of both approaches, as well as some of their computational advantages and limitations. We conclude that, although the image-based approach holds great promise, it has potential pitfalls that may be best overcome by including structural information. Thus, the most viable model of object recognition may be one that incorporates the most appealing aspects of both image-based and structural-description theories.

Original languageEnglish
Pages (from-to)1-20
Number of pages20
JournalCognition
Volume67
Issue number1-2
Publication statusPublished - 1998 Jul 17
Externally publishedYes

Fingerprint

Psychophysics
Neurophysiology
Structural Models
Haplorhini
Retina
Light
psychophysics
neurophysiology
Monkey
Object Recognition
evidence

Keywords

  • Image-based model
  • Object recognition
  • Structural description

ASJC Scopus subject areas

  • Language and Linguistics
  • Cognitive Neuroscience
  • Experimental and Cognitive Psychology
  • Linguistics and Language

Cite this

Image-based object recognition in man, monkey and machine. / Tarr, Michael J.; Bulthoff, Heinrich.

In: Cognition, Vol. 67, No. 1-2, 17.07.1998, p. 1-20.

Research output: Contribution to journalArticle

Tarr, Michael J. ; Bulthoff, Heinrich. / Image-based object recognition in man, monkey and machine. In: Cognition. 1998 ; Vol. 67, No. 1-2. pp. 1-20.
@article{a801e12d180a48b78529ea3d1beba47e,
title = "Image-based object recognition in man, monkey and machine",
abstract = "Theories of visual object recognition must solve the problem of recognizing 3D objects given that perceivers only receive 2D patterns of light on their retinae. Recent findings from human psychophysics, neurophysiology and machine vision provide converging evidence for 'image-based' models in which objects are represented as collections of viewpoint-specific local features. This approach is contrasted with 'structural-description' models in which objects are represented as configurations of 3D volumes or parts. We then review recent behavioral results that address the biological plausibility of both approaches, as well as some of their computational advantages and limitations. We conclude that, although the image-based approach holds great promise, it has potential pitfalls that may be best overcome by including structural information. Thus, the most viable model of object recognition may be one that incorporates the most appealing aspects of both image-based and structural-description theories.",
keywords = "Image-based model, Object recognition, Structural description",
author = "Tarr, {Michael J.} and Heinrich Bulthoff",
year = "1998",
month = "7",
day = "17",
language = "English",
volume = "67",
pages = "1--20",
journal = "Cognition",
issn = "0010-0277",
publisher = "Elsevier",
number = "1-2",

}

TY - JOUR

T1 - Image-based object recognition in man, monkey and machine

AU - Tarr, Michael J.

AU - Bulthoff, Heinrich

PY - 1998/7/17

Y1 - 1998/7/17

N2 - Theories of visual object recognition must solve the problem of recognizing 3D objects given that perceivers only receive 2D patterns of light on their retinae. Recent findings from human psychophysics, neurophysiology and machine vision provide converging evidence for 'image-based' models in which objects are represented as collections of viewpoint-specific local features. This approach is contrasted with 'structural-description' models in which objects are represented as configurations of 3D volumes or parts. We then review recent behavioral results that address the biological plausibility of both approaches, as well as some of their computational advantages and limitations. We conclude that, although the image-based approach holds great promise, it has potential pitfalls that may be best overcome by including structural information. Thus, the most viable model of object recognition may be one that incorporates the most appealing aspects of both image-based and structural-description theories.

AB - Theories of visual object recognition must solve the problem of recognizing 3D objects given that perceivers only receive 2D patterns of light on their retinae. Recent findings from human psychophysics, neurophysiology and machine vision provide converging evidence for 'image-based' models in which objects are represented as collections of viewpoint-specific local features. This approach is contrasted with 'structural-description' models in which objects are represented as configurations of 3D volumes or parts. We then review recent behavioral results that address the biological plausibility of both approaches, as well as some of their computational advantages and limitations. We conclude that, although the image-based approach holds great promise, it has potential pitfalls that may be best overcome by including structural information. Thus, the most viable model of object recognition may be one that incorporates the most appealing aspects of both image-based and structural-description theories.

KW - Image-based model

KW - Object recognition

KW - Structural description

UR - http://www.scopus.com/inward/record.url?scp=0032112199&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0032112199&partnerID=8YFLogxK

M3 - Article

C2 - 9735534

AN - SCOPUS:0032112199

VL - 67

SP - 1

EP - 20

JO - Cognition

JF - Cognition

SN - 0010-0277

IS - 1-2

ER -