The importance of symmetry and virtual views in three-dimensional object recognition

T. Vetter, T. Poggio, Heinrich Bulthoff

Research output: Contribution to journalArticle

107 Citations (Scopus)

Abstract

Background: Human observers can recognize three-dimensional objects seen in novel orientations, even when they have previously seen only a relatively small number of different views of the object. How our visual system does this is a key problem in vision research. Recent theories and experiments suggest that the human visual system might store a relatively small number of sample two-dimensional views of a three-dimensional object, and recognize novel views by a process of interpolation between the stored sample views. These sample views may be collected during a training phase as the visual system familiarizes itself with the object. Results: Here, we investigate whether constraints on the shapes of objects commonly encountered in the real world can reduce the number of training views required for recognition of three-dimensional objects. We are particularly concerned with the constraint of object symmetry. We show that if an object is bilaterally symmetrical, then additional 'virtual views' can automatically be generated from one sample view by symmetry transformations. These virtual views should make it more easy to recognize novel views of a symmetric than an asymmetric object, when a single sample view has been seen. Recognition should be particularly facilitated when the novel views are close to the virtual view. We present psychophysical results that bear out these predictions. Conclusions: Our results show that the human visual system can indeed exploit symmetry to facilitate object recognition, and support the model for object recognition in which a small number of two-dimensional views are remembered and combined to recognize novel views of the same object. These results raise questions about how symmetry is recognized, and symmetry transformations implemented, in real, biological neural networks.

Original languageEnglish
Pages (from-to)18-23
Number of pages6
JournalCurrent Biology
Volume4
Issue number1
Publication statusPublished - 1994 Dec 1
Externally publishedYes

Fingerprint

Object recognition
Interpolation
Neural networks
sampling
Experiments
neural networks
Research
prediction

ASJC Scopus subject areas

  • Agricultural and Biological Sciences(all)

Cite this

The importance of symmetry and virtual views in three-dimensional object recognition. / Vetter, T.; Poggio, T.; Bulthoff, Heinrich.

In: Current Biology, Vol. 4, No. 1, 01.12.1994, p. 18-23.

Research output: Contribution to journalArticle

@article{aaaafdee530f4670a34bf89fc41f3113,
title = "The importance of symmetry and virtual views in three-dimensional object recognition",
abstract = "Background: Human observers can recognize three-dimensional objects seen in novel orientations, even when they have previously seen only a relatively small number of different views of the object. How our visual system does this is a key problem in vision research. Recent theories and experiments suggest that the human visual system might store a relatively small number of sample two-dimensional views of a three-dimensional object, and recognize novel views by a process of interpolation between the stored sample views. These sample views may be collected during a training phase as the visual system familiarizes itself with the object. Results: Here, we investigate whether constraints on the shapes of objects commonly encountered in the real world can reduce the number of training views required for recognition of three-dimensional objects. We are particularly concerned with the constraint of object symmetry. We show that if an object is bilaterally symmetrical, then additional 'virtual views' can automatically be generated from one sample view by symmetry transformations. These virtual views should make it more easy to recognize novel views of a symmetric than an asymmetric object, when a single sample view has been seen. Recognition should be particularly facilitated when the novel views are close to the virtual view. We present psychophysical results that bear out these predictions. Conclusions: Our results show that the human visual system can indeed exploit symmetry to facilitate object recognition, and support the model for object recognition in which a small number of two-dimensional views are remembered and combined to recognize novel views of the same object. These results raise questions about how symmetry is recognized, and symmetry transformations implemented, in real, biological neural networks.",
author = "T. Vetter and T. Poggio and Heinrich Bulthoff",
year = "1994",
month = "12",
day = "1",
language = "English",
volume = "4",
pages = "18--23",
journal = "Current Biology",
issn = "0960-9822",
publisher = "Cell Press",
number = "1",

}

TY - JOUR

T1 - The importance of symmetry and virtual views in three-dimensional object recognition

AU - Vetter, T.

AU - Poggio, T.

AU - Bulthoff, Heinrich

PY - 1994/12/1

Y1 - 1994/12/1

N2 - Background: Human observers can recognize three-dimensional objects seen in novel orientations, even when they have previously seen only a relatively small number of different views of the object. How our visual system does this is a key problem in vision research. Recent theories and experiments suggest that the human visual system might store a relatively small number of sample two-dimensional views of a three-dimensional object, and recognize novel views by a process of interpolation between the stored sample views. These sample views may be collected during a training phase as the visual system familiarizes itself with the object. Results: Here, we investigate whether constraints on the shapes of objects commonly encountered in the real world can reduce the number of training views required for recognition of three-dimensional objects. We are particularly concerned with the constraint of object symmetry. We show that if an object is bilaterally symmetrical, then additional 'virtual views' can automatically be generated from one sample view by symmetry transformations. These virtual views should make it more easy to recognize novel views of a symmetric than an asymmetric object, when a single sample view has been seen. Recognition should be particularly facilitated when the novel views are close to the virtual view. We present psychophysical results that bear out these predictions. Conclusions: Our results show that the human visual system can indeed exploit symmetry to facilitate object recognition, and support the model for object recognition in which a small number of two-dimensional views are remembered and combined to recognize novel views of the same object. These results raise questions about how symmetry is recognized, and symmetry transformations implemented, in real, biological neural networks.

AB - Background: Human observers can recognize three-dimensional objects seen in novel orientations, even when they have previously seen only a relatively small number of different views of the object. How our visual system does this is a key problem in vision research. Recent theories and experiments suggest that the human visual system might store a relatively small number of sample two-dimensional views of a three-dimensional object, and recognize novel views by a process of interpolation between the stored sample views. These sample views may be collected during a training phase as the visual system familiarizes itself with the object. Results: Here, we investigate whether constraints on the shapes of objects commonly encountered in the real world can reduce the number of training views required for recognition of three-dimensional objects. We are particularly concerned with the constraint of object symmetry. We show that if an object is bilaterally symmetrical, then additional 'virtual views' can automatically be generated from one sample view by symmetry transformations. These virtual views should make it more easy to recognize novel views of a symmetric than an asymmetric object, when a single sample view has been seen. Recognition should be particularly facilitated when the novel views are close to the virtual view. We present psychophysical results that bear out these predictions. Conclusions: Our results show that the human visual system can indeed exploit symmetry to facilitate object recognition, and support the model for object recognition in which a small number of two-dimensional views are remembered and combined to recognize novel views of the same object. These results raise questions about how symmetry is recognized, and symmetry transformations implemented, in real, biological neural networks.

UR - http://www.scopus.com/inward/record.url?scp=0028050839&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0028050839&partnerID=8YFLogxK

M3 - Article

VL - 4

SP - 18

EP - 23

JO - Current Biology

JF - Current Biology

SN - 0960-9822

IS - 1

ER -