 |
University of Maryland, College Park, Maryland USA and Army Research Laboratory (ARL) |
 |
|
University of Maryland, College Park, Maryland USA and Army Research Laboratory (ARL) |
|
The DoD Sequoyah Foreign Language Translation Program, managed by the interim Sequoyah Transition Mgt Office (STMO) under PEO IEWS, Ft Monmouth, NJ, is intended to address critical linguist shortfalls in US warfighting and intelligence operations through automated language translation capabilities (speech and text) and to provide document image processing and OCR capabilities for cases when material to be translated is paper or document images. To support unbiased, vendor-neutral assessment of technology candidates prior to field testing and deployment, the STMO has initiated a web-accessible, distributed “Best-of Breed Configurable Active Testbed” (BOBCAT) led and operated by ARL and distributed across NRL and AFRL. Yet to be incorporated into the testbed is the capability to assess OCR and other document image processing (DIP) tools. The STMO as well as the ODNI have tasked ARL with integrating document image (DI) processing assessment into BOBCAT, creatingBOBCAT-DI . BOBCAT-DI will be used to assess a variety of document image processing capabilities and tools, with a focus on Arabic and other Southwest Asian languages. The image processing and analysis metrics and methods developed by UMD/UMIACS, particularly as applied to document images obtained from cameras, scanners, etc., is needed to enable assessments that are reliable, robust, and scientifically defensible.