VML-HD: The historical Arabic documents dataset for recognition systems

Majeed Kassis; Alaa Abdalhaleem; Ahmad Droby; Reem Alaasam; Jihad El-Sana

doi:10.1109/asar.2017.8067751

VML-HD: The historical Arabic documents dataset for recognition systems

Majeed Kassis, Alaa Abdalhaleem, Ahmad Droby, Reem Alaasam, Jihad El-Sana

Department of Computer Science

Research output: Contribution to conference › Paper › peer-review

24 Scopus citations

Abstract

In this paper we present a new database with handwritten Arabic script. It is based on five books written by different writers from the years 1088-1451. We took 680 pages from these five books, and fully annotated them on the sub-word level. For each page we manually applied bounding boxes on the different sub-words and annotated the sequence of characters. It consists of 121,636 sub-word appearances consisted of 244,553 characters out of a vocabulary of 1,731 forms of sub-words. The database is described in detail and is designed for training and testing recognition systems for handwritten Arabic sub-words. This database is available for the purpose of research, and we encourage researchers to develop and test new methods using our database.

Original language	English
Pages	11-14
Number of pages	4
DOIs	https://doi.org/10.1109/asar.2017.8067751
State	Published - 13 Oct 2017
Event	1st IEEE International Workshop on Arabic Script Analysis and Recognition, ASAR 2017 - Nancy, France Duration: 3 Apr 2017 → 5 Apr 2017

Conference

Conference	1st IEEE International Workshop on Arabic Script Analysis and Recognition, ASAR 2017
Country/Territory	France
City	Nancy
Period	3/04/17 → 5/04/17

ASJC Scopus subject areas

Computer Vision and Pattern Recognition
Linguistics and Language
Computer Science Applications

Access to Document

10.1109/asar.2017.8067751

Cite this

@conference{ed02c90a13c44e2fbe123dedeea338c9,

title = "VML-HD: The historical Arabic documents dataset for recognition systems",

abstract = "In this paper we present a new database with handwritten Arabic script. It is based on five books written by different writers from the years 1088-1451. We took 680 pages from these five books, and fully annotated them on the sub-word level. For each page we manually applied bounding boxes on the different sub-words and annotated the sequence of characters. It consists of 121,636 sub-word appearances consisted of 244,553 characters out of a vocabulary of 1,731 forms of sub-words. The database is described in detail and is designed for training and testing recognition systems for handwritten Arabic sub-words. This database is available for the purpose of research, and we encourage researchers to develop and test new methods using our database.",

author = "Majeed Kassis and Alaa Abdalhaleem and Ahmad Droby and Reem Alaasam and Jihad El-Sana",

note = "Funding Information: This research was supported in part by the Lynn and William Frankel Center for Computer Sciences at Ben-Gurion University, Israel, and we{\textquoteright}d like to thank them for their support. We would also like to thank Rami Amsis and Ghassan Nabary for their help in creating the ground truth information for the dataset. We would like also to thank Hussien Othman for proofing and finalizing the ground truth information. Publisher Copyright: {\textcopyright} 2017 IEEE; 1st IEEE International Workshop on Arabic Script Analysis and Recognition, ASAR 2017 ; Conference date: 03-04-2017 Through 05-04-2017",

year = "2017",

month = oct,

day = "13",

doi = "10.1109/asar.2017.8067751",

language = "English",

pages = "11--14",

}

TY - CONF

T1 - VML-HD: The historical Arabic documents dataset for recognition systems

AU - Kassis, Majeed

AU - Abdalhaleem, Alaa

AU - Droby, Ahmad

AU - Alaasam, Reem

AU - El-Sana, Jihad

N1 - Funding Information: This research was supported in part by the Lynn and William Frankel Center for Computer Sciences at Ben-Gurion University, Israel, and we’d like to thank them for their support. We would also like to thank Rami Amsis and Ghassan Nabary for their help in creating the ground truth information for the dataset. We would like also to thank Hussien Othman for proofing and finalizing the ground truth information. Publisher Copyright: © 2017 IEEE

PY - 2017/10/13

Y1 - 2017/10/13

N2 - In this paper we present a new database with handwritten Arabic script. It is based on five books written by different writers from the years 1088-1451. We took 680 pages from these five books, and fully annotated them on the sub-word level. For each page we manually applied bounding boxes on the different sub-words and annotated the sequence of characters. It consists of 121,636 sub-word appearances consisted of 244,553 characters out of a vocabulary of 1,731 forms of sub-words. The database is described in detail and is designed for training and testing recognition systems for handwritten Arabic sub-words. This database is available for the purpose of research, and we encourage researchers to develop and test new methods using our database.

AB - In this paper we present a new database with handwritten Arabic script. It is based on five books written by different writers from the years 1088-1451. We took 680 pages from these five books, and fully annotated them on the sub-word level. For each page we manually applied bounding boxes on the different sub-words and annotated the sequence of characters. It consists of 121,636 sub-word appearances consisted of 244,553 characters out of a vocabulary of 1,731 forms of sub-words. The database is described in detail and is designed for training and testing recognition systems for handwritten Arabic sub-words. This database is available for the purpose of research, and we encourage researchers to develop and test new methods using our database.

UR - http://www.scopus.com/inward/record.url?scp=85045206221&partnerID=8YFLogxK

U2 - 10.1109/asar.2017.8067751

DO - 10.1109/asar.2017.8067751

M3 - Paper

SP - 11

EP - 14

T2 - 1st IEEE International Workshop on Arabic Script Analysis and Recognition, ASAR 2017

Y2 - 3 April 2017 through 5 April 2017

ER -

VML-HD: The historical Arabic documents dataset for recognition systems

Abstract

Conference

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this