Page 326 - Fundamentals of Light Microscopy and Electronic Imaging

RECORD KEEPING DURING IMAGE ACQUISITION AND PROCESSING 309

THE NEED FOR IMAGE PROCESSING

Although our goal is to standardize image-processing operations, our experience tells us
that image processing is an invaluable agent in scientific discovery. Video-enhanced
contrast imaging of microtubules and minute cellular structures, and contrast manipula-
tions to medical X rays and astronomical images are all examples. Here, it is very clear
that nothing can be seen without substantial adjustments to the original image. Further,
two important technological advances in electronic microscope imaging—video
enhancement of image contrast and confocal microscopy—profoundly changed the way
scientists regarded microscope images. These technologies are entirely dependent on
electronic detectors, and both require extensive electronic adjustments and follow-up
computer processing. The images and knowledge gained from these technologies are
impressive. Nevertheless, for the typical microscope image our approach to processing
and image display usually needs to be more conservative.
We would be remiss if we did not emphasize an obvious point. If processing is
done to enhance the visibility of important image features that cannot otherwise be
seen to good advantage, then it is appropriate and acceptable. This is no different from
any other scientific data. Naturally, if you have only one picture showing some piece of
information, then it is not wise to trust it. But if the same feature shows up repeatedly,
then you consider it trustworthy even if it is difficult to see. In such situations, image
processing might be the only way to display image details and describe an observation
to others.

VARYING PROCESSING STANDARDS

Images used as data in scientific publications are distinct from the cover images of sci-
entific journals or the enhanced images used to highlight scientific presentations. These
two categories of images are both important, but serve different purposes and are pre-
pared using different standards. Ideally, figures of scientific merit must represent a spec-
imen objectively and completely. The fundamental tenet of scientific publishing—that
the reader should be able to repeat observations based on the details given in the arti-
cle—should be the goal. Images on journal covers serve a different purpose. Here the
rules are relaxed, and artistic manipulation of the image may be important to emphasize
a particular feature, although the cover legend should acknowledge this point. Experi-
ence in image processing suggests it is best to act conservatively. You will have the con-
fidence that you have prepared an objective record based on defined parameters for
acquisition and processing.

RECORD KEEPING DURING IMAGE ACQUISITION
AND PROCESSING

Good image processing begins with keeping good records.

• Save raw and processed image files separately. Before processing an image, make
a copy of the original raw image and save the processed image as a separate file.
When preparing figures for publication, keep notes about the processing steps
applied to each image. Try to standardize your operations so that images are treated