Classifying surface texture while simultaneously estimating illumination direction

M. Chantler, M. Petrou, A. Penirsche, M. Schmidt, G. McGunnigle

Research output: Contribution to journalArticle

Abstract

We propose a novel classifier that both classifies surface texture and simultaneously estimates the unknown illumination conditions. A new formal model of the dependency of texture features on lighting direction is developed which shows that their mean vectors are trigonometric functions of the illuminations' tilt and slant angles. This is used to develop a probabilistic description of feature behaviour which forms the basis of the new classifier. Given a feature set from an image of an unknown texture captured under unknown illumination conditions the algorithm first estimates the most likely illumination direction for each possible texture class. These estimates are used to calculate the class likelihoods and the classification is made accordingly. The ability of the classifier to estimate illuminant direction, and to assign the correct class, was tested on 55 real texture samples in two stages. The classifier was able to accurately estimate both the tilt and the slant angles of the light source for the majority of textures and gave a 98% classification rate.

Original languageEnglish
Pages (from-to)83-96
Number of pages14
JournalInternational Journal of Computer Vision
Volume62
Issue number1-2
DOIs
Publication statusPublished - Apr 2005

Fingerprint

Textures
Lighting
Classifiers
Light sources

Cite this

Chantler, M. ; Petrou, M. ; Penirsche, A. ; Schmidt, M. ; McGunnigle, G. / Classifying surface texture while simultaneously estimating illumination direction. In: International Journal of Computer Vision. 2005 ; Vol. 62, No. 1-2. pp. 83-96.
@article{ce05694f549a426ca0c03b07d454d726,
title = "Classifying surface texture while simultaneously estimating illumination direction",
abstract = "We propose a novel classifier that both classifies surface texture and simultaneously estimates the unknown illumination conditions. A new formal model of the dependency of texture features on lighting direction is developed which shows that their mean vectors are trigonometric functions of the illuminations' tilt and slant angles. This is used to develop a probabilistic description of feature behaviour which forms the basis of the new classifier. Given a feature set from an image of an unknown texture captured under unknown illumination conditions the algorithm first estimates the most likely illumination direction for each possible texture class. These estimates are used to calculate the class likelihoods and the classification is made accordingly. The ability of the classifier to estimate illuminant direction, and to assign the correct class, was tested on 55 real texture samples in two stages. The classifier was able to accurately estimate both the tilt and the slant angles of the light source for the majority of textures and gave a 98{\%} classification rate.",
author = "M. Chantler and M. Petrou and A. Penirsche and M. Schmidt and G. McGunnigle",
year = "2005",
month = "4",
doi = "10.1023/B:VISI.0000046590.98379.19",
language = "English",
volume = "62",
pages = "83--96",
journal = "International Journal of Computer Vision",
issn = "0920-5691",
publisher = "Springer",
number = "1-2",

}

Classifying surface texture while simultaneously estimating illumination direction. / Chantler, M.; Petrou, M.; Penirsche, A.; Schmidt, M.; McGunnigle, G.

In: International Journal of Computer Vision, Vol. 62, No. 1-2, 04.2005, p. 83-96.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Classifying surface texture while simultaneously estimating illumination direction

AU - Chantler, M.

AU - Petrou, M.

AU - Penirsche, A.

AU - Schmidt, M.

AU - McGunnigle, G.

PY - 2005/4

Y1 - 2005/4

N2 - We propose a novel classifier that both classifies surface texture and simultaneously estimates the unknown illumination conditions. A new formal model of the dependency of texture features on lighting direction is developed which shows that their mean vectors are trigonometric functions of the illuminations' tilt and slant angles. This is used to develop a probabilistic description of feature behaviour which forms the basis of the new classifier. Given a feature set from an image of an unknown texture captured under unknown illumination conditions the algorithm first estimates the most likely illumination direction for each possible texture class. These estimates are used to calculate the class likelihoods and the classification is made accordingly. The ability of the classifier to estimate illuminant direction, and to assign the correct class, was tested on 55 real texture samples in two stages. The classifier was able to accurately estimate both the tilt and the slant angles of the light source for the majority of textures and gave a 98% classification rate.

AB - We propose a novel classifier that both classifies surface texture and simultaneously estimates the unknown illumination conditions. A new formal model of the dependency of texture features on lighting direction is developed which shows that their mean vectors are trigonometric functions of the illuminations' tilt and slant angles. This is used to develop a probabilistic description of feature behaviour which forms the basis of the new classifier. Given a feature set from an image of an unknown texture captured under unknown illumination conditions the algorithm first estimates the most likely illumination direction for each possible texture class. These estimates are used to calculate the class likelihoods and the classification is made accordingly. The ability of the classifier to estimate illuminant direction, and to assign the correct class, was tested on 55 real texture samples in two stages. The classifier was able to accurately estimate both the tilt and the slant angles of the light source for the majority of textures and gave a 98% classification rate.

U2 - 10.1023/B:VISI.0000046590.98379.19

DO - 10.1023/B:VISI.0000046590.98379.19

M3 - Article

VL - 62

SP - 83

EP - 96

JO - International Journal of Computer Vision

JF - International Journal of Computer Vision

SN - 0920-5691

IS - 1-2

ER -