Home > Technical Notes

Technical Notes

The goal in digitizing material from Justine's stamp collection is to provide access to it. The following information provides details about the digitization and ingestion (uploading and cataloging) process.

Justine's personal stamp collection has more than 1,500 stamps in seven topic categories, Authors, Books & Libraries, Computers, Flowers, Printing, and Telecomm. The first phase was to digitize, upload, and add metadata for 167 stamps from the "Books & Libraries" category. The stamps are organized and displayed on the web using Omeka, a web-publishing system. Upcoming phrases will cover stamps in the remaining categories.

Scanning and file management

The scanner used to digitize the stamps comes with the Brother MFC 8480DN, a multi-function printer with a flatbed scanner and scans up to 600 dpi (dots per inch). The printer comes with scanning software with support for JPEG, TIFF, PNG, JPEG 2000 (JP2), GIF, BMP, and PDF formats. The printer is primarily used for home office use.

Each stamp is scanned in color for an 8 bit, JPEG 2000 image at 600 dpi and saved using the filename that identifies its topic, country, and Scott number (see Filenames below). A copy of each JP2 file is created for backup.

To prepare the images for web display, each JP2 file is imported into Adobe Photoshop, smudges removed, and straightened (with the Clone Stamp and Measure It tools). After the images are cleaned, the "Save for web" function reduces the image to 72 dpi, and saves in JPG format.

Filenames

To identify each digitized stamp image's file, the syntax for the file's name is based on the stamp's topic, country, and Scott number in the format: topic_country_scottnumber. For example, the Danish stamp with the Scott Number 525 from the Books & Libraries category is identified as "bl_denmark_s525." The filename ends with the appropriate filetype extension, .jp2 or .jpg.

Metadata

Justine keeps descriptive data about each stamp in a Microsoft Access 2000 database, along with a thumbnail image of the stamp. (Justine used Access' "create data access page" function which can presumably be used to create web pages. But efforts to use this function were unsuccessful.) This data is exported in CVS and imported to a spreadsheet application. The data is normalized to match Dublin Core element fields, for example, collection (aka category), country, and year. Contributor, source, and subject information are added to the sheet.

The following Dublin Core elements in Omeka are used.

  • Contributor - the person adding the stamps to the resource.
  • Country - the area where the stamp was issued.
  • Coverage - the country's geographic region.
  • Creator - the designer(s) of the stamp.
  • Date - when the stamp was issued.
  • Description - information about the stamp.
  • Rights - the stamp's copyright information.
  • Source - the source of the stamps. Although only one source (Justine's album) is currently used in this collection, stamps from another collection may be considered in the future.
  • Subject - the key topic(s) of the stamp.
  • Title - the name given to the stamp, either by the Post Office, or identified via research.

For metadata that is standard in stamp collecting, the following fields were added in Omeka under Still Image type elements.

  • Condition - the original condition of the stamp, mint or used.
  • Denomination - the face value of the stamp.
  • Physical Dimensions - approximate size of the stamp.
  • Scott Number - the stamp's identifying number in the Scott Standard Postage Stamp Catalogue.

Subject headings are selected from the Library of Congress Thesaurus for Graphic Materials 1 (TGM 1), to provide a controlled vocabulary for the collection. Names of the stamp's Country are "preferred" terms according to the Getty Thesaurus of Geographic Names® Online.

All fields are searchable. Additional metadata is added to the stamp's record in Omeka as it becomes available.

Uploading to Omeka

Omeka's CVS Import and Dropbox plugins are used to upload the metadata and image files. CVS Import creates a new record for each row of data in the import CVS file. After the images are uploaded, each one is manually linked to the appropriate record. During this process, Omeka automatically creates three images in different dimensions: a full-size, a thumbnail size, and a square thumbnail size for the Browse pages. The dimensions are preset in Omeka's settings.

General Workflow

  1. Export metadata in Access database in CVS to Excel. Create a tracker sheet from the data.
  2. Scan the stamp, save it as JP2, and log it in sheet.
  3. After all stamps in category are digitized, a copy is made as a backup.
  4. Each stamp is inspected in Adobe Photoshop, cleaned and straightened where necessary. When this is completed, it is noted in tracker sheet.
  5. Normalize the spreadsheet to match Dublin Core elements and fill in as many fields as possible. For example, a Contributor column is added and filled in.
  6. Upload the cvs file.
  7. Upload the images in batches.
  8. Link images to their records.
  9. Add new metadata to records as it becomes available.