[fitsbits] Re: FITS for Image Archiving?

[Steve Willner]

In article <39411470.B507760B at emory.edu>, Patrick Durusau
<pdurusau at emory.edu> writes: 
> The images I am particularly interested in would be of printed materials
> but there are other members of the group that will be scanning items
> such as historical movie posters and the like.
> 
> Has anyone used FITS as an archival format for non-astronomical data?
> More to the point, has anyone used FITS to store scanned images of
> printed materials?

Just a few comments from a non-expert, if it's not too late.

The strong points of FITS for image storage are, in my opinion, 1)
extensive header information stored with the image, and 2) strict
preservation of the data values of all pixels.  FITS is, I am told,
used in medical imaging as well as astronomy.  I can well imagine a
doctor needing to know, for example, whether one body part is twice
as opaque or three times as opaque as another part, and only a format
that preserves data values can provide this information.

It is not so obvious to me that the strengths of FITS are the ones
needed for scans of printed materials.  Header information could be
provided in a README file accompanying each image file.  Headers will
likely be produced by human typing anyway, not by an automated
acquisition device, and they are likely to be used by humans, not by
processing software.  Also, I don't think most scanners will give any
close relation between data numbers and fundamental properties of the
original object.  Data values will depend on the calibration of the
scanner and may not be linear, so preserving the exact values doesn't
seem too important.  You would, I expect, like to preserve
_differences_, so that parts of the object that have the same
appearance end up with the same data value, and parts having
different appearance have different values, but I think most image
formats that aren't highly compressed will do that.

On the other hand, I don't think there's anything terribly wrong with
using FITS.  It preserves all the information, and it's portable.  It
just may not be as efficient or as easy to use for your purposes as
other formats.

-- 
Steve Willner            Phone 617-495-7123     swillner at cfa.harvard.edu
Cambridge, MA 02138 USA                 
(Please email your reply if you want to be sure I see it; include a
valid Reply-To address to receive an acknowledgement.  Commercial
email may be sent to your ISP.)