Several sets of lecture notes and tutorials
are available for download or on-line access:
- A 90 page, 23 MByte pdf file that includes the numerous
image processing and measurement examples
from my annual one-day workshops presented at the MSA meetings. Emphasis is on
the best workflow and the
procedures that can be applied to each step. The examples use Photoshop plus some plug-ins.
- The Molecular Expressions web site hosted by Mike Davidson's group at Florida State University has an
on-line set of
my lecture notes enhanced by dozens of interactive Java applets, covering many of the topics that I
present in a typical one-day tutorial.
- I am frequently asked to provide a brief (one-day or half-day) introductory session on Photoshop, before
extended hands-on workshops on image processing and analysis. This is intended to help those unfamiliar
with the Adobe Photoshop environment learn about the program's basic functions, general appearance, and
those capabilities which it offers that are appropriate for scientific imaging. A 29-page tutorial
can be downloaded as a 12MB pdf file. A 2.5 MB
zip file containing the images used in the tutorial is also provided.
- Another common concern is the choice of a camera appropriate for use on a light microscope. I personally
use a wide variety of cameras, ranging from specialized cameras intended for that application to general
purpose digital cameras including single-lens-reflex models with adaptors, and even consumer cameras that look
through the microsope eyepiece (which have the advantage of portability). Some of the factors to consider
when choosing a camera are described in this brief writeup.
- Lossy compression of images, most often using the JPEG algorithm, is simply unacceptable if they are to be
used for scientific or forensic purposes. An article illustrates some of
the types of artefacts that are caused, both in photographs and in thresholded binary images intended for measurement.
- A 150 page tutorial showing the step-by-step use of
Photoshop and Fovea Pro, which was revised
and rewritten for the most recent (version 4.0) release of the software by Reindeer Graphics. The
entire tutorial can be downloaded or accessed
on-line as a series of pdf files.
- An 11 page condensed summary (containing many of the illustrations) from my invited lecture on forensic
image processing and analysis applications,
presented at the North Carolina State University
Forensic Sciences Symposium, can be downloaded as a pdf file. The
entire lecture is also viewable on-line for those who wish to
watch it.
Some downloadable Photoshop plug-ins and other goodies:
- The original Fovea Pro 4.0 plugins for Macintosh were written for the PowerPC processor and can only be used
with the newer generation
Intel computers using Rosetta. A complete set of the plug-ins in native Intel code, and full information on using
them, is available
here. They are limited to 32-bit versions of Photoshop (CS5 and older).
- The widely used (and frequently abused) unsharp mask procedure works by subtracting a Gaussian blurred copy
of the image (which removes important detail) from the original. The difference, which is the detail, is then
added back to the original to enhance the visibility of steps, lines and edges. However, the process also
creates haloes around edges, and produces different results in bright and dark areas of the image. An
alternate method that uses rank based (morphological) operators rather than the Gaussian blur avoids these
problems and generally produces superior results. A Photoshop plug-in for
Windows or Macintosh can be downloaded that
implements this method. Once placed in your filters folder it will be listed as Filters > IP*Xtras > RankSharp
and will run with no dialog. The accompanying text file explains the parameters used and how they can be adjusted.
At several meetings in the US and overseas, I have been asked to present a plenary lecture on
the characteristics of our human visual system and how it interacts with the images that are
obtained from various kinds of microscopy (and other scientific images). The basic message is that
humans are not very good observers, that our vision system ignores a lot of information, that
having names and labels for recognized features is very important, and that we often think we see
what we expect to see. After one recent presentation, urged on by several of the people at the
conference, I have prepared a written version of the paper which was serialized in the
Proceedings of the Royal Microscopy Society. It is about 25,000 words long, has more than 80
illustrations, and can be downloaded as a 6 MB Acrobat
(pdf) file. In the presentation I include a few Quicktime movies, which are shown in the pdf file
as single still frames. This is probably sufficient for the reader to understand the message, but
if anyone wants the actual 600K of *.mov files they can also be downloaded
(Figures 23, 27,
35 and 36).
Whether images are being used for scientific or forensic purposes, it is vital to understand how
processing affects the perception or measurement of the scene. The ethics of what constitutes proper
methods for image processing is something that must be considered. Here is my statement on this subject.
Page updated May 27, 2016
For further info:
