ICDAR 2017 Competition on Historical Document Writer Identification (Historical-WI)


Writer identification is a behavioral handwriting-based recognition problem which proceeds by matching unknown handwritings against a database of samples with known authorship. From the document image analysis scope, writer identification can be defined as the retrieval of handwritten samples of the same writer from a database using a handwritten sample as a graphical query. The large number of recent publications as well as the organization of several competitions, proves that writer identification is a very active and promising area of research. The identification of the writer of a handwritten document has a wide variety of applications. For example, analysis of handwritten documents has great bearing on the criminal justice systems. Other application areas include security, financial activity, forensic analysis and access control.


Following the successful organization of ICDAR2011 Writer Identification Contest, ICFHR2012 Writer Identification Contest Challenge 1: Latin Documents and ICDAR2013 Competition on Writer Identification, a competition for writer identification tecniques is organized in the framework of ICDAR2017, in order to record recent advances in the field of writer identification. This competition deals with the identification of writers in historical handwritten documents. The task is to generate a ranking of the images stored in the database according to the similarity of the handwriting. The ranking for each page (most similar on the first place) is then submitted to the competition website and the identification rate and mean average precision is then calculated.

The dataset used for this competition consists of 3600 document images, which have been written by 720 different writers. Each writer has contributed 5 documents. The images are provided as color images, but also a biniarized version is provided. You can download the dataset here:

The files do not contain any information about the writer. You have the generate a csv file like discribed here and submit it to ScripNet. The best result for each group are revealed after the competition has ended.

A trainings dataset with similar images (and GT) can be downloaded here: It consists out 394 writers, with 3 pages each. The document images should have the same properties like in the dataset used for the competition. We want to stress out, that no writer of this set has any page in the dataset for the competition. It is just to get an impression of the performance of your algorithm, but note that since the dataset has less pages and less writers the performance on both sets is not directly connected.

We invite all researchers in the field of writer identification to register on the ScriptNet platform and participate in ICDAR 2017 Competition on Historical Document Writer Identification (Historical-WI). The description of the methods and the evaluation scores will be presented during a dedicated ICDAR2017 contest session. A report on the competition will be published in the ICDAR2017 conference proceedings.

Note: You may participate in this contest even if you do not plan to attend the ICDAR 2017 conference.

To contact the organizers, write an email to fiel@caa.tuwien.ac.at

News

Website online Dataset released change of the submission date

Important Dates

Release of the Dataset: 17.03.2017

Submission open 23.06.2017

Submission Deadline: 05.07.2017
12.07.2017





Organizers







Stefan Fiel

[TU Wien, Computer Vision Lab] 

Stefan Fiel works as research assistant at the Computer Vision Lab, TU Wien.

Basilis Gatos

[NCSR Demokritos / IIT / CIL] 

Researcher at the Institute of Informatics and Telecommunications of the National Center for Scientific Research "Demokritos", Athens, Greece.

Georgios Louloudis

[NCSR Demokritos / IIT / CIL] 

Research Associate at the Institute of Informatics and Telecommunications of the National Center for Scientific Research "Demokritos", Athens, Greece.

Nikolaos Stamatopoulos

[NCSR Demokritos / IIT / CIL] 

Research Associate at the Institute of Informatics and Telecommunications of the National Center for Scientific Research "Demokritos", Athens, Greece.

Markus Diem

[TU Wien, Computer Vision Lab] 

Markus Diem is a senior scientist at the Computer Vision Lab, TU Wien, Austria. His research interests are Cultural Heritage Applications and Document Analysis.

Florian Kleber

[TU Wien, Computer Vision Lab] 

Florian Kleber is currently a senior scientist at the Computer Vision Lab, Institute for Computer Aided Automation, TU Wien, Austria. His research interests are Cultural Heritage Applications and Document Analysis Applications.

Vincent Christlein

[FAU Erlangen-N├╝rnberg, Pattern Recognition Lab] 

Vincent Christlein is a PhD student at the Pattern Recognition Lab, FAU Erlangen-Nuremberg.