Learning perceptual texture similarity and relative attributes from computational features

Jianwen Lou; Lin Qi; Junyu Dong; Hui Yu; Guoqiang Zhong

doi:10.1109/IJCNN.2016.7727516

Learning perceptual texture similarity and relative attributes from computational features

Jianwen Lou, Lin Qi, Junyu Dong, Hui Yu, Guoqiang Zhong

University of Portsmouth Ageing Network

Ocean University of China

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

379 Downloads (Pure)

Abstract

Previous work has shown that perceptual texture similarity and relative attributes cannot be well described by computational features. In this paper, we propose to predict human’s visual perception of texture images by learning a nonlinear
mapping from computational feature space to perceptual space. Hand-crafted features and deep features, which were successfully applied in texture classification tasks, were extracted and used to train Random Forest and rankSVM models against perceptual data from psychophysical experiments. Three texture datasets were used to test our proposed method and the experiments show that the predictions of such learnt models are in high correlation with human’s results.

Original language	English
Title of host publication	2016 International Joint Conference on Neural Networks (IJCNN)
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	2540-2546
Number of pages	7
ISBN (Electronic)	978-1-5090-0620-5
ISBN (Print)	978-1-5090-0621-2
DOIs	https://doi.org/10.1109/IJCNN.2016.7727516
Publication status	Published - 3 Nov 2016
Event	2016 IEEE World Congress on Computational Intelligence - Vancouver, Canada Duration: 25 Jul 2016 → 29 Jul 2016

Publication series

Name
ISSN (Electronic)	2161-4407

Conference

Conference	2016 IEEE World Congress on Computational Intelligence
Abbreviated title	IEEE WCCI
Country/Territory	Canada
City	Vancouver
Period	25/07/16 → 29/07/16

Keywords

feature extraction
correlation
semantics
somputational modeling
neural networks
manifolds
symmetric matrices

Access to Document

10.1109/IJCNN.2016.7727516

Learning Perceptual Texture Similarity_IJCNN16
© 2016 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.
Accepted author manuscript (Post-print), 766 KB

Cite this

@inproceedings{70d248352b3d4875bc1ecd7aabb3d39c,

title = "Learning perceptual texture similarity and relative attributes from computational features",

abstract = "Previous work has shown that perceptual texture similarity and relative attributes cannot be well described by computational features. In this paper, we propose to predict human{\textquoteright}s visual perception of texture images by learning a nonlinearmapping from computational feature space to perceptual space. Hand-crafted features and deep features, which were successfully applied in texture classification tasks, were extracted and used to train Random Forest and rankSVM models against perceptual data from psychophysical experiments. Three texture datasets were used to test our proposed method and the experiments show that the predictions of such learnt models are in high correlation with human{\textquoteright}s results.",

keywords = "feature extraction, correlation, semantics, somputational modeling, neural networks, manifolds, symmetric matrices",

author = "Jianwen Lou and Lin Qi and Junyu Dong and Hui Yu and Guoqiang Zhong",

year = "2016",

month = nov,

day = "3",

doi = "10.1109/IJCNN.2016.7727516",

language = "English",

isbn = "978-1-5090-0621-2",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

pages = "2540--2546",

booktitle = "2016 International Joint Conference on Neural Networks (IJCNN)",

address = "United States",

note = "2016 IEEE World Congress on Computational Intelligence, IEEE WCCI ; Conference date: 25-07-2016 Through 29-07-2016",

}

Lou, J, Qi, L, Dong, J, Yu, H & Zhong, G 2016, Learning perceptual texture similarity and relative attributes from computational features. in 2016 International Joint Conference on Neural Networks (IJCNN). Institute of Electrical and Electronics Engineers Inc., pp. 2540-2546, 2016 IEEE World Congress on Computational Intelligence, Vancouver, Canada, 25/07/16. https://doi.org/10.1109/IJCNN.2016.7727516

TY - GEN

T1 - Learning perceptual texture similarity and relative attributes from computational features

AU - Lou, Jianwen

AU - Qi, Lin

AU - Dong, Junyu

AU - Yu, Hui

AU - Zhong, Guoqiang

PY - 2016/11/3

Y1 - 2016/11/3

N2 - Previous work has shown that perceptual texture similarity and relative attributes cannot be well described by computational features. In this paper, we propose to predict human’s visual perception of texture images by learning a nonlinearmapping from computational feature space to perceptual space. Hand-crafted features and deep features, which were successfully applied in texture classification tasks, were extracted and used to train Random Forest and rankSVM models against perceptual data from psychophysical experiments. Three texture datasets were used to test our proposed method and the experiments show that the predictions of such learnt models are in high correlation with human’s results.

AB - Previous work has shown that perceptual texture similarity and relative attributes cannot be well described by computational features. In this paper, we propose to predict human’s visual perception of texture images by learning a nonlinearmapping from computational feature space to perceptual space. Hand-crafted features and deep features, which were successfully applied in texture classification tasks, were extracted and used to train Random Forest and rankSVM models against perceptual data from psychophysical experiments. Three texture datasets were used to test our proposed method and the experiments show that the predictions of such learnt models are in high correlation with human’s results.

KW - feature extraction

KW - correlation

KW - semantics

KW - somputational modeling

KW - neural networks

KW - manifolds

KW - symmetric matrices

U2 - 10.1109/IJCNN.2016.7727516

DO - 10.1109/IJCNN.2016.7727516

M3 - Conference contribution

SN - 978-1-5090-0621-2

SP - 2540

EP - 2546

BT - 2016 International Joint Conference on Neural Networks (IJCNN)

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 2016 IEEE World Congress on Computational Intelligence

Y2 - 25 July 2016 through 29 July 2016

ER -

Learning perceptual texture similarity and relative attributes from computational features

Abstract

Publication series

Conference

Keywords

Access to Document

Fingerprint

Cite this