Image Banks

[Peter Rauch]

On Thu, 10 Dec 1998, John D. Oswald wrote:
>    The Department of Entomology at Texas AM University will soon be
> developing a local "image bank" to support a variety of web-based projects
> currently on-line and under development by various faculty and staff.

John, I suppose that your image bank could contain a variety of kinds of
images, such as photographic, satellite, GIS maps/overlays, graphics,
documents, etc. Were you wanting to include these various types, or were
you thinking of limiting it to photographic (digital) images perhaps?

If photographic, what photographical media are the sources for the image
bank's digital versions? E.g., exclusively 35mm color transp., b&w/color
6x6 negs, b&w/color prints, ...?

What about the size of the image bank? Hundreds, thousands, tens of
1000's of source images?

> We wish to
> develop a locally centralized "image bank" within the department that will
> include the image files themselves, and data associated with the images
> (e.g., content, credits, etc.).

Ahh, "content". Photographic images often contain a great diversity of
"content".  The various contents (subject matter) often are of interest
to a variety of disciplines, and for a variety of uses. Enriching an
image database by including not only the ability to add an
indefinite number of "content" values, but to allow image
users/viewers who are expert in one or another aspect of the image's
content to add those values, on the fly, will help in the enrichment
process.

> .. Some of the issues of interest to us include: (1)
> the efficient storage and organization of image files,

How about adding "efficient digital image capture" to your list of
issues? Perhaps the more costly effort will be digitizing the existing
collection(s) of images. It's a process that you may not want to repeat,
so getting it right the first go-around seems prudent. Depending on what
kinds of source media (see above) you plan to digitize, the process
would vary.

One of the important points to keep in mind with photographic media
(esp. the common 35mm color transparency) is the "resolution" at which
the source images are digitized; you'll want to scan as much of the
original "data" to be found in the source image as you think users will
find useful. Because there is great detail to be found in many photo
images, there are compelling arguments for digitizing the sources at
very high resolution.

Have you already digitized your collections of source images?

> (2) the types of
> "associated" data that "should" be collected ...

Yes. This refers to metadata in general, and back to my comments above
about "content" as well.

No answers, John. Just some particular points and issues to keep in
mind. Perhaps if folks had a little more detail on what kinds of media,
numbers, etc., some further suggestions and some answers might be easier
to offer.
Peter