The Historical AND Digital Brown

1
THE HISTORICAL AND DIGITAL BrownWhite
Turning reels and paper into a CONTENTdm
collection with article level access
Christine Guenther Product Manger, Digital
Services Bethlehem, PAformerly OCLC
Preservation Service Center
2
The Roadmap
Selection
ACCESSQUALITY
METADATAQUALITY
IMAGEQUALITY
3
Workflow Choices
Lehigh U Project
4
(No Transcript)
5
Evaluation of source material
Start with FILM or digitize ORIGINAL?
? Print Master? Preservation grade film?
Budget gt Scan from existing film
? Poor filming? Only poor copies? (color
content) gtScan direct
6
Hybrid approach scan good film, use originals
for the other volumes
112 years, 114 volumes, about 55,000
pagesProduction 2007/2008.
7
Scanning Original BW volumes

• Zeutschel 7000 bookscanner
• 400 dpi capture
• 8-bit grayscale
• Bound volumes vs. gutter text loss

8
Scanning Preservation Microfilm

• NextScan Eclipse
• 35mm roll film scanner
• Same specifications
• Pro/Con Scan rate high, but post-scan image
  enhancements necessary (deskew, crop etc)

9
Quality Assurance!

• Bethlehem Digital QA team
• View every image for proper capture and
  completeness
• Collect issue date and page sequence data for all
  pages
• Deal with naming irregularities typical for
  student newspapers or dupes on film

10
From pixels to content
11
Beyond OCR

• Challenge OCR is cost-effective, but only as
  accurate as text quality appearance in the source
  file
• Simple OCR is not reliable for recognizing
  specific metadata such as Issue Date and blind
  for document structure (articles)
• Used metadata elements similar to NDNP guidelines
  (including LCCN, geographic coverage, etc.)
• Advanced Search goes beyond full text
  search!Offering access points for discovery.

12
Metadata collection - Results
13
Content conversion

• CCS docWORKS software
• CCS developed the schema in cooperation with 12
  EU and US libraries during the EU-funded METAe
  project.
• ALTO Analyzed Layout and Text Object
• The Library of Congress chose CCS's ALTO-schema
  for the National Digital Newspaper Program (NDNP)

14
METS/ALTO

• METS Contains all digital preservation data
  bibliographical, administrative, technical,
  structural ONE PER ISSUE
• Passport for digital preservation
• ALTO Contains layout information and OCR results
  each word is mapped to a specific location in
  an image ONE PER PAGE
• Can also include article level information

15
Summary The data package per page
access

• Best Image in standard format
• SUSTAIN
• REPROCESS(?)

• Well suited for oversize content
• QUICK ACCESSw/ DETAIL

• De-facto standard
• PRINT
• CONTAINED

• SearchabilityXML
• FULL TEXT

CCS docWORKS
16
Extra feature Segmentation

• Layout analysis
• Manual correction
• Article jumps
• Headline correction

17
Production floor in India
18
Headline correction
19
Online presentation system CONTENTdm
20
CONTENTdm Collection Building

• Import METS/ALTO JPEG2000 PDF Compound
  Objects (one per issue)
• Troubleshooting
• Quality Assurance

21
Thank you!

• For further information
• Christine Guenther
• Backstage Library Works9 South Commerce
  WayBethlehem, PA 180171-800-773
  7222guenthec_at_oclc.org
• Lara Henry
• Sales Representative
• lara_at_bslw.com
• 1-800-288-1265
• www.bslw.com