International Workshop on Document Image Analysis for Libraries Announced

A two-day workshop on document image analysis for libraries will be held January 23-24, 2004 at Xerox's Palo Alto Research Center in California. DIAL2004 will bring together Digital Library (DL) and Document Image Analysis (DIA) researchers, practitioners, and users who are interested in new technologies that assist the integration of imaged documents within DLs so that (ideally) everything that can be done with symbolically encoded data can also be done with scanned hardcopy documents. In an increasingly digital world, vast legacy (& modern) collections of irreplaceable paper documents are threatened with neglect and irrelevance unless difficult DIA tasks can be automated: searching document images, presenting them legibly, navigating within and among them, etc. This workshop will attempt to describe the state of the art and identify urgent open problems. More broadly, the workshop is designed to promote closer cooperation between the DIA and DL communities in exploring fundamental capabilities that will allow information systems to operate with equal effectiveness across all media types and formats, including paper documents and other human-legible but non-digital media, in multiple languages, and from many historical periods.

The event will include podium presentations, keynote talks, panel discussions, all presented in a single track, in addition to breakout working group sessions. The goal is 100% participation, i.e., every attendee will give a presentation. On-site printed proceedings and a post-proceedings book and/or journal special issue are planned. The event starts immediately following the IS&T/SPIE 16th Annual Symposium on Electronic Imaging: Science & Technology, to be held Jan. 18-22 in San Jose, CA.

Topics of interest include, but are not limited to:

• Document Image Analysis (DIA) technology for Digital Libraries (DLs);
• Challenging DL open problems requiring new DIA research strategies;
• End-user requirements for document images provided via DLs;
• Case studies of DLs that serve document images well;
• Imaging & compression standards for document preservation & analysis;
• Automatic quality control during document image capture;
• Content & metadata extraction, recognition, tagging, linking, etc.;
• Parallel tagging of images, transcripts, and other document layers;
• Information extraction from images of tables, graphs, math, etc.;
• Searching/querying, retrieval, summarizing/condensing of doc images;
• Presentation & legibility of document images on GUIs, eBooks, PDAs;
• On-line & web-based navigation within/among document images;
• Personal & interactive DLs: e.g. capture, correction, reading;
• Historical/archival DIA; original-medium quality challenges;
• Inter- & multi-national DLs: e.g. languages, scripts, translations;
• Guaranteeing authenticity of document images; rights management;
• Citation and editorial control of image-based data;
• File formats & representations for document images & analysis results;
• Multimedia document analysis, including audio & video;
• Proposals for DIA/DL database collection, truthing, & benchmarking;
• Government DL projects (e.g. The Newton Project, American Memory); and
• Critical surveys of the state of the art of DLs & DIA.

We invite two classes of submissions:

Regular Paper (for researchers, professional practitioners): up to 30 pages, refereed, published, long presentations
Abstract of Remarks (for businesses, end-users, students): 1-5 pages, not refereed, unpublished, brief presentations

Deadlines for regular papers:

October 15 Submission
November 20 Acceptance
December 20 Camera-ready copy

Deadlines for abstracts of remarks:

December 1 Submission
December 15 Acceptance

For further information, see www.cedar.buffalo.edu/DIAL2004, or send inquiries to : dial2004@cedar.buffalo.edu or by mail to the CEDAR Secretariat, Univ. at Buffalo, 520 Lee Entrance, #202, Amherst, NY 14228