NSF DLI All-Projects Meeting Stratford-upon-Avon The Digital Atheneum New

NSF/DLI All-Projects Meeting Stratford-upon-Avon The Digital Atheneum: New Techniques for Restoring, Searching and Editing Digital Collections W. Brent Seales Jim Griffioen Kevin S. Kiernan University of Kentucky

NSF/DLI All-Projects Meeting Stratford-upon-Avon The Digital Atheneum • Project Goals – Develop new techniques for digital preservation, restoration, and presentation of damaged manuscripts • University of Kentucky Principal Investigators: – Jim Griffioen and Brent Seales (Computer Science) – Kevin Kiernan (English) • Partners – The British Library – IBM

NSF/DLI All-Projects Meeting Stratford-upon-Avon Damaged Manuscripts • Problems: – Requires extensive editorial involvement – Manuscript is fragile – Single images mean little without metadata – Need new restoration tools and techniques

NSF/DLI All-Projects Meeting Stratford-upon-Avon Research Areas • Digitization techniques – – 3 D shape many types of illumination • Content-based searching and access – – Extensible/flexible editorial tools Distributed execution of tools and services • Restoration techniques – Mosaicing – Digital flattening/shape visualization

NSF/DLI All-Projects Meeting Stratford-upon-Avon Illumination Techniques Visible light, Ultraviolet light, Laser scanning, Fiber-optic back-lighting

NSF/DLI All-Projects Meeting Stratford-upon-Avon Three-Dimensional Scanning • Digital photography is used to provide patrons electronic facsimiles of library materials • 2 D images do not provide shape information – Insufficient in providing a realistic “look and feel” of non-planar materials • We present a technique that augments current digitization setups to acquire 3 D information

NSF/DLI All-Projects Meeting Stratford-upon-Avon Digitization of Materials Special Collections: Culturally significant, unique, damaged, restricted materials Patrons are passionate about the look and feel of the original Want more than just “textual content” It is typical to provide the patron images alone (no 3 D information)

NSF/DLI All-Projects Meeting Stratford-upon-Avon Digitizing “Manuscripts” • We address digitization of manuscripts and similar objects – Objects are ill-suited for flat-bed scanners – Photographed with digital camera

NSF/DLI All-Projects Meeting Stratford-upon-Avon Typical Digitization Setup • • • High-resolution digital camera Camera positioned above the “artifact” Image is captured fsjerjwee ael Worjt ileke sldkfj Mwiee kerlk

NSF/DLI All-Projects Meeting Stratford-upon-Avon Multiple Lighting Conditions • Common to illuminate same object under several lighting conditions Natural fsjerjwee ael Worjt ileke sldkfj Mwiee kerlk

NSF/DLI All-Projects Meeting Stratford-upon-Avon Multiple Lighting Conditions • Common to illuminate same object under several lighting conditions White fsjerjwee ael Worjt ileke sldkfj Mwiee kerlkerlk

NSF/DLI All-Projects Meeting Stratford-upon-Avon Multiple Lighting Conditions • Common to illuminate same object under several lighting conditions UV fsjerjwee ael Worjt ileke sldkfj Mwiee kerlkerlk

NSF/DLI All-Projects Meeting Stratford-upon-Avon Limitation of 2 D imagery When imaging non-flat objects Objects can appear crinkled or warped Difficult to disambiguate if the warping is part of the original or an artifact of the object’s shape 2 D imagery is limited in what it can convey to the patron

NSF/DLI All-Projects Meeting Stratford-upon-Avon Beyond 2 D images With non-planar materials 2 D images can prove insufficient in providing “look and feel” 3 D representations provide better electronic facsimiles Goal: provide a textured 3 D model to patron

NSF/DLI All-Projects Meeting Stratford-upon-Avon Our Goal • Provide a 3 D acquisition system that can be integrated into current digitization setup – Curator decides when to acquire “ 3 D” – Otherwise operates as usual • No special-purpose equipment – Uses equipment already available – Minimum overhead

3 D Acquisition setup for library materials NSF/DLI All-Projects Meeting Stratford-upon-Avon • Introduce a light projector – Projector mathematics behaves same as camera • Projector and camera are used to triangulate 3 D points on the artifact’s surface • A 3 D model can be acquired (and 2 D image) y z x

NSF/DLI All-Projects Meeting Stratford-upon-Avon Current Implementation • Prototype system – Uses standard NTSC camera (low-resolution) – XGA light projector (1024 x 768) • Mirror is used to direct the light onto the materials – Camera and projector are controlled from the same computer (provides synchronization) – Produces 3 D model with several textures

NSF/DLI All-Projects Meeting Stratford-upon-Avon Current Implementation • Acquisition Timing – 5 minutes to calibrate – 1 -3 minutes per object • Can scan several objects after calibration • Scan Density – typically less than 0. 5 mm along projector X&Y • Accuracy – 0. 25 mm Results with low-resolution camera

Results NSF/DLI All-Projects Meeting Stratford-upon-Avon Materials from U. Kentucky’s Special Collection Non-planar materials 13 th century wax seals Washington Coin Book binding (inset)

NSF/DLI All-Projects Meeting Stratford-upon-Avon Editing and Searching Content • Data components to manage: – Images – Text (transcription, edition) – Glossaries – Metadata – Embedded content

NSF/DLI All-Projects Meeting Stratford-upon-Avon Toolkit Objectives • Integrated interface – visualization – editing/creation – content-based search • • Adaptable to new problem domains Extensible Distributed Collaborative (multi-user access)

NSF/DLI All-Projects Meeting Stratford-upon-Avon Distributed Service-base Architecture • The set of services can be easily modified and extended • Base Services – SQL-style DB service, Image and Transcript viewer/editor service, Correspondence service, Display service, Access/Security service • Example Extended Services – Navigation services, Template-based Image Search services, Model-based Search services, Enhancement and Restoration services, Glossary services, Annotation service, . . . • Current implementation based on Java/Jini allows services to be run locally or at the server

NSF/DLI All-Projects Meeting Stratford-upon-Avon The Editor’s Tool - Example

NSF/DLI All-Projects Meeting Stratford-upon-Avon The Editor’s Tool (1)

NSF/DLI All-Projects Meeting Stratford-upon-Avon The Editor’s Tool (2)

NSF/DLI All-Projects Meeting Stratford-upon-Avon The Editor’s Tool (3)

NSF/DLI All-Projects Meeting Stratford-upon-Avon The Editor’s Tool (4)

NSF/DLI All-Projects Meeting Stratford-upon-Avon The Editor’s Tool (5)

NSF/DLI All-Projects Meeting Stratford-upon-Avon The Editor’s Tool (6)

NSF/DLI All-Projects Meeting Stratford-upon-Avon The Editor’s Tool (7)

NSF/DLI All-Projects Meeting Stratford-upon-Avon The Editor’s Tool (8)

NSF/DLI All-Projects Meeting Stratford-upon-Avon The Editor’s Tool (9)

NSF/DLI All-Projects Meeting Stratford-upon-Avon The Editor’s Tool (10)

NSF/DLI All-Projects Meeting Stratford-upon-Avon The Editor’s Tool (11)

NSF/DLI All-Projects Meeting Stratford-upon-Avon Future Work • • Digitize British Library Cottonian collection Enhance content-based search functionality Integrate 3 D data Develop restoration techniques

NSF/DLI All-Projects Meeting Stratford-upon-Avon The Digital Atheneum: New Techniques for Restoring, Searching and Editing Digital Collections www. digitalatheneum. org University of Kentucky