Offline Writer Identification Using K-Adjacent Segments

Author

Jain, Rajiv ; Doermann, David

Author_Institution

Univ. of Maryland, College Park, MD, USA

fYear

2011

fDate

18-21 Sept. 2011

Firstpage

769

Lastpage

773

Abstract

This paper presents a method for performing offline writer identification by using K-adjacent segment (KAS) features in a bag-of-features framework to model a user´s handwriting. This approach achieves a top 1 recognition rate of 93% on the benchmark IAM English handwriting dataset, which outperforms current state of the art features. Results further demonstrate that identification performance improves as the number of training samples increase, and additionally, that the performance of the KAS features extend to Arabic handwriting found in the MADCAT dataset.

Keywords

document image processing; handwritten character recognition; natural language processing; Arabic handwriting; IAM English handwriting dataset; K-adjacent segment; KAS features; MADCAT dataset; bag-of-features framework; offline writer identification; user handwriting; Accuracy; Feature extraction; Hidden Markov models; Image segmentation; Testing; Training; Vectors; Codebook; Document Forensics; Handwriting; K-Adjacent Segments; Local Features; Writer Identification;

fLanguage

English

Publisher

ieee

Conference_Titel

Document Analysis and Recognition (ICDAR), 2011 International Conference on

Conference_Location

Beijing

ISSN

1520-5363

Print_ISBN

978-1-4577-1350-7

Electronic_ISBN

1520-5363

Type

conf

DOI

10.1109/ICDAR.2011.159

Filename

6065415