Synthetic faces, face cubes, and the geometry of face space

Hugh R. Wilson, Gunter Loffler, Frances Wilkinson

Research output: Contribution to journalArticle

Abstract

To simplify the study of visual face processing, we introduce a novel class of synthetic face stimuli based upon 37 measurements (head shape, feature locations, etc.) extracted from individual face photographs in both frontal and 20° side views. Synthetic faces are bandpass filtered optimally for face perception and include both line and edge information. Pilot experiments establish that subjects are extremely accurate in matching a synthetic face with the original grayscale photograph, even across views. To determine the perceptual metric of face space, we introduce face cubes in which the geometric differences between any faces in a four-dimensional face subspace can be precisely determined. Experiments on face discrimination using face cubes establish the metric of synthetic face space as locally Euclidean, with discrimination thresholds representing 4–6% total geometric variation (as a percent of mean head radius) between faces. Discrimination thresholds are lowest for face cubes constructed around the average face, thus indicating that the mean face for each gender represents a natural origin for face space. Finally, synthetic faces exhibit a pronounced inversion effect for 20° side views and a characteristic “Thatcher effect” for inverted front views. Synthetic faces and face cubes thus provide a useful new quantitative approach to the study of face perception and face space.

Original languageEnglish
JournalVision Research
DOIs
Publication statusPublished - 1 Dec 2002

Fingerprint

Head
Discrimination (Psychology)
Facial Recognition

Keywords

  • pattern discrimination
  • face perception
  • stimulus geometry
  • form vision
  • face space

Cite this

@article{d0b6fc46f9784eae89b5f5da15ba6f53,
title = "Synthetic faces, face cubes, and the geometry of face space",
abstract = "To simplify the study of visual face processing, we introduce a novel class of synthetic face stimuli based upon 37 measurements (head shape, feature locations, etc.) extracted from individual face photographs in both frontal and 20° side views. Synthetic faces are bandpass filtered optimally for face perception and include both line and edge information. Pilot experiments establish that subjects are extremely accurate in matching a synthetic face with the original grayscale photograph, even across views. To determine the perceptual metric of face space, we introduce face cubes in which the geometric differences between any faces in a four-dimensional face subspace can be precisely determined. Experiments on face discrimination using face cubes establish the metric of synthetic face space as locally Euclidean, with discrimination thresholds representing 4–6{\%} total geometric variation (as a percent of mean head radius) between faces. Discrimination thresholds are lowest for face cubes constructed around the average face, thus indicating that the mean face for each gender represents a natural origin for face space. Finally, synthetic faces exhibit a pronounced inversion effect for 20° side views and a characteristic “Thatcher effect” for inverted front views. Synthetic faces and face cubes thus provide a useful new quantitative approach to the study of face perception and face space.",
keywords = "pattern discrimination, face perception, stimulus geometry, form vision, face space",
author = "Wilson, {Hugh R.} and Gunter Loffler and Frances Wilkinson",
note = "Originally published in: Vision Research (2002), 42 (27), pp.2909-2923.",
year = "2002",
month = "12",
day = "1",
doi = "10.1016/S0042-6989(02)00362-0",
language = "English",
journal = "Vision Research",
issn = "0042-6989",
publisher = "Elsevier Ltd",

}

Synthetic faces, face cubes, and the geometry of face space. / Wilson, Hugh R.; Loffler, Gunter; Wilkinson, Frances.

In: Vision Research, 01.12.2002.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Synthetic faces, face cubes, and the geometry of face space

AU - Wilson, Hugh R.

AU - Loffler, Gunter

AU - Wilkinson, Frances

N1 - Originally published in: Vision Research (2002), 42 (27), pp.2909-2923.

PY - 2002/12/1

Y1 - 2002/12/1

N2 - To simplify the study of visual face processing, we introduce a novel class of synthetic face stimuli based upon 37 measurements (head shape, feature locations, etc.) extracted from individual face photographs in both frontal and 20° side views. Synthetic faces are bandpass filtered optimally for face perception and include both line and edge information. Pilot experiments establish that subjects are extremely accurate in matching a synthetic face with the original grayscale photograph, even across views. To determine the perceptual metric of face space, we introduce face cubes in which the geometric differences between any faces in a four-dimensional face subspace can be precisely determined. Experiments on face discrimination using face cubes establish the metric of synthetic face space as locally Euclidean, with discrimination thresholds representing 4–6% total geometric variation (as a percent of mean head radius) between faces. Discrimination thresholds are lowest for face cubes constructed around the average face, thus indicating that the mean face for each gender represents a natural origin for face space. Finally, synthetic faces exhibit a pronounced inversion effect for 20° side views and a characteristic “Thatcher effect” for inverted front views. Synthetic faces and face cubes thus provide a useful new quantitative approach to the study of face perception and face space.

AB - To simplify the study of visual face processing, we introduce a novel class of synthetic face stimuli based upon 37 measurements (head shape, feature locations, etc.) extracted from individual face photographs in both frontal and 20° side views. Synthetic faces are bandpass filtered optimally for face perception and include both line and edge information. Pilot experiments establish that subjects are extremely accurate in matching a synthetic face with the original grayscale photograph, even across views. To determine the perceptual metric of face space, we introduce face cubes in which the geometric differences between any faces in a four-dimensional face subspace can be precisely determined. Experiments on face discrimination using face cubes establish the metric of synthetic face space as locally Euclidean, with discrimination thresholds representing 4–6% total geometric variation (as a percent of mean head radius) between faces. Discrimination thresholds are lowest for face cubes constructed around the average face, thus indicating that the mean face for each gender represents a natural origin for face space. Finally, synthetic faces exhibit a pronounced inversion effect for 20° side views and a characteristic “Thatcher effect” for inverted front views. Synthetic faces and face cubes thus provide a useful new quantitative approach to the study of face perception and face space.

KW - pattern discrimination

KW - face perception

KW - stimulus geometry

KW - form vision

KW - face space

U2 - 10.1016/S0042-6989(02)00362-0

DO - 10.1016/S0042-6989(02)00362-0

M3 - Article

JO - Vision Research

JF - Vision Research

SN - 0042-6989

ER -