Second International Workshop on Multilingual OCR

25^th July, 2009

In conjunction with ICDAR 2009

Universitat Autònoma de Barcelona, Catalonia - Spain

*NEW* Workshop Program is now online. *NEW*

Instructions for submission of camera-ready copies of accepted papers can be found on the Submissions page.

Registration for the workshop will be handled by the ICDAR 2009 registration system.

Visa Assistance and Letters: Individuals requiring an official letter of invitation in order to attend the Workshop may send e-mail to the ICDAR2009 Secretariat (icdar2009@ultramarevents.com). This procedure is designed to assist participants who need to obtain a visa or permission to attend the conference, but does not cover registration fees or other expenses. The invitation does not imply provision of financial or other support. If you don't receive a confirmation of your request within a week, re-send your inquiry immediately. More information can be found at the Travel Information section of the ICDAR 2009 web site.

MOCR 2009 invites the submission of original, previously unpublished work in any of the areas of interest to the multilingual OCR community. Submissions could also include preliminary experimental results based on new approaches, descriptions of newly created open data sets, proposals for competitions and extended work highlighting recent results that were too late for the ICDAR 2009deadline.

The workshop will explore methodologies for multilingual document analysis systems with particular focus on OCR. The scope of Multilingual OCR is defined to include systems that are capable of reading more than one language in the same document, as well as one-language-per-document systems that can be easily retargeted to new languages. The workshop will provide a forum for technical discussions on three important themes:

• progress in the last decade in multilingual OCR
• adaptation and repurposing of proven methods for multilingual OCR
• hard open research problems and promising new approaches

The workshop will focus on the following and related topics

• Evaluation Methodologies: metrics, standards, ground truth; benchmark datasets
• Proven Methodologies: HMMs, Script Identification, Layout Analysis, Reading Order
• Mixed languages: Techniques applicable/retargetable to multiple languages/scripts; documents containing multiple languages (single-line, document level, e.g. dictionary),
• Newer languages/ Scripts: Techniques for dealing with problems of scripts for which OCR technology has not matured
• Different levels: Language vs. Scripts; Machine-print& Handwriting; Image & e-text
• Document Analysis: Mathematical formulas; tables; videotext; abbreviations; annotations
• Degraded Documents: Colloquialism, dialect, headline, image degradation, connectedness