Electronic Work Flow and Digital Asset Management, Part I
Vol. 30 No. 1 2004
William Willner RPB, FBPA
This paper, the first of two parts, is concerned primarily with the procedures established for the movement of digital assets within the department of Biomedical Communications, Wake Forest University School of Medicine. A more thorough discussion of digital asset management as it relates to electronic work flow will be the subject of Part II.
Storing and retrieving images has always been a frustrating and time-consuming endeavor. In the past, slides, negatives, and prints had to be stored in a fashion that would ensure no change to their physical characteristics over time. Today, most images are created electronically and stored as files on a computer system. While the risk of physical changes and damage is near non-existent with electronic assets, storing and retrieving image files, in general, is not much easier than it was with their film-based counterparts. A standardized system of strictly-adhered-to procedures is the only way to make certain that the image files stored today may be easily retrieved tomorrow.
The system used at Wake Forest School of Medicine can be broken into four interrelated tasks:
This system currently has about 50,000 records spread over nine databases, representing 34 GB of on-line, and 111 GB of off-line stored image files. We define asset management as those tasks required to label and classify images so that they and any associated information are easily retrievable. We use Cumulus 5.5 (www.canto.com) and Rename 1.4a (www.1-4a.com (www.1-4a.com) for this purpose. The end product is a series of catalogs, each of which contains a thumbnail of the original photograph, image annotation information, and user-applied categories that relate to image content.
Electronic work flow is defined as movement of files from initial capture to final archive. We use Cumulus 5.5 and Adobe Photoshop (www.adobe.com) for these tasks. Our department is server centric. All images are saved on a server. Only programs reside on local machines. Substituting “folder” for “volume” would allow a non-server-based department to implement a similar scheme.
A brief overview of file movement and server structure follows:
Raw Volume Files
After a photographer returns from an assignment, images are renamed, copied to the server, cataloged, and captioned. An assignment is not considered complete until all these tasks are performed. A detailed description follows.
Figure 2. Cumulus Caption field is shown with caption added. Cumulus writes this information back to the file’s IPTC header. In addition to caption information, it is also possible to save information related to credits, bylines, caption writer, copyright, date/time, and location in the IPTC header. We use only the caption field.
Figure 3. Cumulus is opened over Photoshop. Dragging the thumbnail from Cumulus to Photoshop opens the file in that application. The Cumulus thumbnails are large enough to perform a rough edit. Final edit decisions are made from the files opened in Photoshop.
Figure 5. This screen grab is of the image archive database. The right portion shows the Cumulus asset information screen. Notice that the database is set to capture IPTC captioning information. The asset information screen also displays those categories that are unique to this image. These categories are chosen from the complete category list shown on the left (see also in Table 2).
Files are renamed using Rename 1.4a according to the following scheme:
Alpha Alpha, Date, Sequential Number (AAmmddyy##)
The two-letter alpha represents a broad categorization of the image’s subject. All of our categories are listed in Table I. This naming scheme ensures that all file names are unique. Also, and just as important, the file name provides some information about the image. For example, an image named PE100202107 is an environmental portrait shot on October 2, 2002.
The specific categories one uses are not important. What is important, however, is to have enough categories to ensure some separation of images, but not so many that it becomes difficult to decide into which category an image belongs.
Copying to, and Retrieving, from the Server
All camera files, except portraits and medical images, are saved on the “Raw Files 2004” server volume (There also were volumes labeled Raw Files 2003, Raw Files 2002, etc.). Absolutely everything is saved. We do not allow any “in camera” editing. If an image is on a card, it will be copied to the server.
The raw file volume is subdivided into a series of folders named Raw##-mm-yy (e.g, Raw01-01-04, Raw02-01-04, Raw03-02-04). This sequence ensures uniqueness of the folder name. We reset the folder number (##) to 01 on January 1. Each folder is limited to 650 MB. If the images on a card force the folder size above 650 MB, a new folder is started (Raw ##+1-mm-yy). We try not to split an assignment between folders.
The images in the raw file folders are written monthly to CD, and the images on the server are deleted. The CDs are labeled with the corresponding folder name (e.g., Raw##-mm-yy). This is usually done two months after the images are taken; (May’s images, for instance, are written to CD in July. This is probably one of the most important image database management tasks we perform. Regardless of the amount of hard drive storage you have, eventually that space will be filled. A plan for the removal of on-line images is vital.
Cataloging (the First Cumulus Catalog)
All camera files from the server volume “Raw Files 2004” are cataloged in a single Cumulus catalog named “Raw Files 2004” (We also have catalogs “Raw Files 2003 and Raw Files 2002”). We have forced Cumulus to make larger than normal, 512 x 512 pixel, thumbnails (Catalog - Properties - General - Thumbnail Size 512). It is easy to do a first-image edit with thumbnails of this size (Figure 1).
After cataloging, file annotation information is added to the Cumulus caption field (Figure 2). Cumulus 5.5 writes the captioning information back to the file's IPTC (International Press Telecommunications Council) header. The captioning information actually becomes a part of the file, travels with the file, and is accessible through Photoshop (File - File Info) (Figure 3). Consequently, images sent electronically to editors or designers will have captioning information available. This is a major advantage of Cumulus. Since the captioning information travels with the image file, the possibility of incorrectly labeling image content is significantly reduced.
The caption is a searchable Cumulus field. This information, combined with the categories contained in the file name, allows for an easily searched database.
Working Volume Files
The raw file catalog is opened over Photoshop (Figure 4). The Cumulus thumbnails are large enough to allow for a rough image edit. Dragging a thumbnail from Cumulus to Photoshop opens the original file (“asset” in Cumulus parlance). Final image selections are made from those files opened in Photoshop.
As soon as the edit is complete, selected images are saved as .psd files to the working volume of the server into one of the following folders:.
Standard image correction routines are applied to the saved photographs. These images reside here until they have been used for their primary purposes. Usually, the images are used within a month.
Archive Volume Files
The “to be cataloged” folders are swept monthly. All the images that have been used for their primary purposes are opened in Photoshop, flattened, and saved as JPEGs in a corresponding set of folders (image archive, infinity archive, group photo archive) on the archive volume of the server. Each of these folders has a matching Cumulus catalog, which is also updated monthly. These catalogs are set to capture the caption information in the image’s IPTC header. Only the image archive catalog requires any additional work. These photographs receive a more complete categorization (Figure 5). Table 2 provides a list of our categories.
As department archivist, I am the only one with write rights to the archive volume. All others within the department have read-only rights. This limitation prevents accidental overwriting of archived images.
Since a relatively high percentage of photographs taken today are electronic, a different strategy must be used for storage and retrieval of these images than that which was used for their silver-based counterparts. Archiving software, such as Cumulus, allows for the creation of databases to help facilitate the management of these digital assets.
Table 1. Image Categories
Table 2. Cumulus Categories
Will Willner (email@example.com)
holds degrees from Case Western Reserve University, Clark University,
and the Rochester Institute of Technology. He is a faculty member and
Coordinator of Photography at Wake Forest University School of Medicine.
He also is an instructor of photography at the North Carolina School of
the Arts. Willner is a Registered Biological Photographer and Fellow of
the BioCommunications Association.