Off-line writer identification using multi-scale local binary patterns and SR-KDA

E. Khalifa; Somaya Al-Maadeed; Muhammad Tahir; Fouad Khelifi; Ahmed Bouridane

doi:10.1109/ICM.2013.6734983

Off-line writer identification using multi-scale local binary patterns and SR-KDA

E. Khalifa, Somaya Al-Maadeed, Muhammad Tahir, Fouad Khelifi, Ahmed Bouridane

Research output: Chapter in Book/Report/Conference proceeding › Chapter › peer-review

1 Citation (Scopus)

Abstract

Writer identification is becoming an increasingly important research topic especially in forensic and biometric applications. This paper presents a novel method for performing offline write identification by using multi-scale local binary patterns histogram (MLBPH) features. The proposed feature (MLBPH) when combined with edge-hinge based feature achieves a top 1 recognition rate of 92% on the benchmark IAM English handwriting dataset, outperforming current state of the art features. Further, kernel discriminant analysis using spectral regression (SR-KDA) is introduced as dimensionality reduction technique to avoid the overfitting problem associated with using multi-scale data.

Original language	English
Title of host publication	2013 25th International Conference on Microelectronics (ICM)
Place of Publication	Piscataway, NJ
Publisher	IEEE
Pages	1-4
ISBN (Print)	978-1-4799-3569-7
DOIs	https://doi.org/10.1109/ICM.2013.6734983
Publication status	Published - 2013

Keywords

Writer Identification
kernel discriminant analysis
multi-scale local binary patterns

Access to Document

10.1109/ICM.2013.6734983

Cite this

@inbook{829a55771e574f7aba55d8669f6d9486,

title = "Off-line writer identification using multi-scale local binary patterns and SR-KDA",

abstract = "Writer identification is becoming an increasingly important research topic especially in forensic and biometric applications. This paper presents a novel method for performing offline write identification by using multi-scale local binary patterns histogram (MLBPH) features. The proposed feature (MLBPH) when combined with edge-hinge based feature achieves a top 1 recognition rate of 92% on the benchmark IAM English handwriting dataset, outperforming current state of the art features. Further, kernel discriminant analysis using spectral regression (SR-KDA) is introduced as dimensionality reduction technique to avoid the overfitting problem associated with using multi-scale data.",

keywords = "Writer Identification, kernel discriminant analysis, multi-scale local binary patterns",

author = "E. Khalifa and Somaya Al-Maadeed and Muhammad Tahir and Fouad Khelifi and Ahmed Bouridane",

year = "2013",

doi = "10.1109/ICM.2013.6734983",

language = "English",

isbn = "978-1-4799-3569-7",

pages = "1--4",

booktitle = "2013 25th International Conference on Microelectronics (ICM)",

publisher = "IEEE",

address = "United States",

}

TY - CHAP

T1 - Off-line writer identification using multi-scale local binary patterns and SR-KDA

AU - Khalifa, E.

AU - Al-Maadeed, Somaya

AU - Tahir, Muhammad

AU - Khelifi, Fouad

AU - Bouridane, Ahmed

PY - 2013

Y1 - 2013

N2 - Writer identification is becoming an increasingly important research topic especially in forensic and biometric applications. This paper presents a novel method for performing offline write identification by using multi-scale local binary patterns histogram (MLBPH) features. The proposed feature (MLBPH) when combined with edge-hinge based feature achieves a top 1 recognition rate of 92% on the benchmark IAM English handwriting dataset, outperforming current state of the art features. Further, kernel discriminant analysis using spectral regression (SR-KDA) is introduced as dimensionality reduction technique to avoid the overfitting problem associated with using multi-scale data.

AB - Writer identification is becoming an increasingly important research topic especially in forensic and biometric applications. This paper presents a novel method for performing offline write identification by using multi-scale local binary patterns histogram (MLBPH) features. The proposed feature (MLBPH) when combined with edge-hinge based feature achieves a top 1 recognition rate of 92% on the benchmark IAM English handwriting dataset, outperforming current state of the art features. Further, kernel discriminant analysis using spectral regression (SR-KDA) is introduced as dimensionality reduction technique to avoid the overfitting problem associated with using multi-scale data.

KW - Writer Identification

KW - kernel discriminant analysis

KW - multi-scale local binary patterns

U2 - 10.1109/ICM.2013.6734983

DO - 10.1109/ICM.2013.6734983

M3 - Chapter

SN - 978-1-4799-3569-7

SP - 1

EP - 4

BT - 2013 25th International Conference on Microelectronics (ICM)

PB - IEEE

CY - Piscataway, NJ

ER -

Off-line writer identification using multi-scale local binary patterns and SR-KDA

Abstract

Keywords

Access to Document

Fingerprint

Cite this