Perceptual Texture Similarity Estimation: An Evaluation of Computational Features

Xinghui Dong, Junyu Dong, Michael John Chantler

Research output: Contribution to journalArticle

Abstract

Estimation of texture similarity is fundamental to many material recognition tasks. This study uses fine-grained human perceptual similarity ground-truth to provide a comprehensive evaluation of 51 texture feature sets. We conduct two types of evaluation and both show that these features do not estimate similarity well when compared against human agreement rates, but that performances are improved when the features are combined using a Random Forest. Using a simple two-stage statistical model we show that few of the features capture long-range aperiodic relationships. We perform two psychophysical experiments which indicate that long-range interactions do provide humans with important cues for estimating texture similarity. This motivates an extension of the study to include Convolutional Neural Networks (CNNs) as they enable arbitrary features of large spatial extent to be learnt. Our conclusions derived from the use of two pre-trained CNNs are: that the large spatial extent exploited by the networks' top convolutional and first fully-connected layers, together with the use of large numbers of filters, confers significant advantage for estimation of perceptual texture similarity.
Original languageEnglish
JournalIEEE Transactions on Pattern Analysis and Machine Intelligence
Early online date7 Jan 2020
DOIs
Publication statusE-pub ahead of print - 7 Jan 2020

Fingerprint

Textures
Neural networks
Experiments

Cite this

@article{ac504ada8c244b17a5f56682c63b1bf8,
title = "Perceptual Texture Similarity Estimation: An Evaluation of Computational Features",
abstract = "Estimation of texture similarity is fundamental to many material recognition tasks. This study uses fine-grained human perceptual similarity ground-truth to provide a comprehensive evaluation of 51 texture feature sets. We conduct two types of evaluation and both show that these features do not estimate similarity well when compared against human agreement rates, but that performances are improved when the features are combined using a Random Forest. Using a simple two-stage statistical model we show that few of the features capture long-range aperiodic relationships. We perform two psychophysical experiments which indicate that long-range interactions do provide humans with important cues for estimating texture similarity. This motivates an extension of the study to include Convolutional Neural Networks (CNNs) as they enable arbitrary features of large spatial extent to be learnt. Our conclusions derived from the use of two pre-trained CNNs are: that the large spatial extent exploited by the networks' top convolutional and first fully-connected layers, together with the use of large numbers of filters, confers significant advantage for estimation of perceptual texture similarity.",
author = "Xinghui Dong and Junyu Dong and Chantler, {Michael John}",
year = "2020",
month = "1",
day = "7",
doi = "10.1109/TPAMI.2020.2964533",
language = "English",
journal = "IEEE Transactions on Pattern Analysis and Machine Intelligence",
issn = "0162-8828",
publisher = "IEEE",

}

TY - JOUR

T1 - Perceptual Texture Similarity Estimation: An Evaluation of Computational Features

AU - Dong, Xinghui

AU - Dong, Junyu

AU - Chantler, Michael John

PY - 2020/1/7

Y1 - 2020/1/7

N2 - Estimation of texture similarity is fundamental to many material recognition tasks. This study uses fine-grained human perceptual similarity ground-truth to provide a comprehensive evaluation of 51 texture feature sets. We conduct two types of evaluation and both show that these features do not estimate similarity well when compared against human agreement rates, but that performances are improved when the features are combined using a Random Forest. Using a simple two-stage statistical model we show that few of the features capture long-range aperiodic relationships. We perform two psychophysical experiments which indicate that long-range interactions do provide humans with important cues for estimating texture similarity. This motivates an extension of the study to include Convolutional Neural Networks (CNNs) as they enable arbitrary features of large spatial extent to be learnt. Our conclusions derived from the use of two pre-trained CNNs are: that the large spatial extent exploited by the networks' top convolutional and first fully-connected layers, together with the use of large numbers of filters, confers significant advantage for estimation of perceptual texture similarity.

AB - Estimation of texture similarity is fundamental to many material recognition tasks. This study uses fine-grained human perceptual similarity ground-truth to provide a comprehensive evaluation of 51 texture feature sets. We conduct two types of evaluation and both show that these features do not estimate similarity well when compared against human agreement rates, but that performances are improved when the features are combined using a Random Forest. Using a simple two-stage statistical model we show that few of the features capture long-range aperiodic relationships. We perform two psychophysical experiments which indicate that long-range interactions do provide humans with important cues for estimating texture similarity. This motivates an extension of the study to include Convolutional Neural Networks (CNNs) as they enable arbitrary features of large spatial extent to be learnt. Our conclusions derived from the use of two pre-trained CNNs are: that the large spatial extent exploited by the networks' top convolutional and first fully-connected layers, together with the use of large numbers of filters, confers significant advantage for estimation of perceptual texture similarity.

U2 - 10.1109/TPAMI.2020.2964533

DO - 10.1109/TPAMI.2020.2964533

M3 - Article

JO - IEEE Transactions on Pattern Analysis and Machine Intelligence

JF - IEEE Transactions on Pattern Analysis and Machine Intelligence

SN - 0162-8828

ER -