Image classification: A novel texture signature approach

Wenda He; Reyer Zwiggelaar

doi:10.1109/ICIP.2010.5654026

Image classification: A novel texture signature approach

Wenda He, Reyer Zwiggelaar

Department of Computer Science

Research output: Chapter in Book/Report/Conference proceeding › Conference Proceeding (ISBN)

Abstract

In this paper we present a novel image classification methodology based on texture signature. The approach consists of four distinct steps: 1) feature extraction from texture images without using any prior knowledge (e.g. viewpoint, illumination condition); 2) textures are modelled as texture signatures; 3) model selection and reduction is used to remove noise and outliers; 4) texture image classification using Columbia-Utrecht (CUReT) texture database. Classification performance was 91% accuracy for all 61 materials (2806 images) present in the CUReT database. The results are compared with texton based classification and effects due to various parameter settings are discussed.

Original language	English
Title of host publication	2010 IEEE International Conference on Image Processing
Publisher	IEEE Press
Pages	2725-2728
Number of pages	4
ISBN (Electronic)	978-1-4244-7994-8
ISBN (Print)	978-1-4244-7992-4
DOIs	https://doi.org/10.1109/ICIP.2010.5654026
Publication status	Published - 26 Sept 2010
Event	2010 17th IEEE International Conference on Image Processing - Hong Kong, China Duration: 26 Sept 2010 → 29 Sept 2010

Conference

Conference	2010 17th IEEE International Conference on Image Processing
Country/Territory	China
City	Hong Kong
Period	26 Sept 2010 → 29 Sept 2010

Keywords

Image Classification
Texture
Signature

Access to Document

10.1109/ICIP.2010.5654026

Permanent link

Persistent link

Cite this

@inproceedings{52200555bdc344e1bdc6c9f2646dc778,

title = "Image classification: A novel texture signature approach ",

abstract = "In this paper we present a novel image classification methodology based on texture signature. The approach consists of four distinct steps: 1) feature extraction from texture images without using any prior knowledge (e.g. viewpoint, illumination condition); 2) textures are modelled as texture signatures; 3) model selection and reduction is used to remove noise and outliers; 4) texture image classification using Columbia-Utrecht (CUReT) texture database. Classification performance was 91\% accuracy for all 61 materials (2806 images) present in the CUReT database. The results are compared with texton based classification and effects due to various parameter settings are discussed.",

keywords = "Image Classification, Texture, Signature",

author = "Wenda He and Reyer Zwiggelaar",

year = "2010",

month = sep,

day = "26",

doi = "10.1109/ICIP.2010.5654026",

language = "English",

isbn = "978-1-4244-7992-4",

pages = "2725--2728",

booktitle = "2010 IEEE International Conference on Image Processing",

publisher = "IEEE Press",

address = "United States of America",

note = "2010 17th IEEE International Conference on Image Processing ; Conference date: 26-09-2010 Through 29-09-2010",

}

TY - GEN

T1 - Image classification

T2 - 2010 17th IEEE International Conference on Image Processing

AU - He, Wenda

AU - Zwiggelaar, Reyer

PY - 2010/9/26

Y1 - 2010/9/26

N2 - In this paper we present a novel image classification methodology based on texture signature. The approach consists of four distinct steps: 1) feature extraction from texture images without using any prior knowledge (e.g. viewpoint, illumination condition); 2) textures are modelled as texture signatures; 3) model selection and reduction is used to remove noise and outliers; 4) texture image classification using Columbia-Utrecht (CUReT) texture database. Classification performance was 91% accuracy for all 61 materials (2806 images) present in the CUReT database. The results are compared with texton based classification and effects due to various parameter settings are discussed.

AB - In this paper we present a novel image classification methodology based on texture signature. The approach consists of four distinct steps: 1) feature extraction from texture images without using any prior knowledge (e.g. viewpoint, illumination condition); 2) textures are modelled as texture signatures; 3) model selection and reduction is used to remove noise and outliers; 4) texture image classification using Columbia-Utrecht (CUReT) texture database. Classification performance was 91% accuracy for all 61 materials (2806 images) present in the CUReT database. The results are compared with texton based classification and effects due to various parameter settings are discussed.

KW - Image Classification

KW - Texture

KW - Signature

U2 - 10.1109/ICIP.2010.5654026

DO - 10.1109/ICIP.2010.5654026

M3 - Conference Proceeding (ISBN)

SN - 978-1-4244-7992-4

SP - 2725

EP - 2728

BT - 2010 IEEE International Conference on Image Processing

PB - IEEE Press

Y2 - 26 September 2010 through 29 September 2010

ER -

Image classification: A novel texture signature approach

Abstract

Conference

Keywords

Access to Document

Permanent link

Fingerprint

Cite this