Some time ago, I issued the following query:
"I am looking for software to measure grain size of thin
sputtered films deposited onto silicon test wafers. The films we need
to measure are either aluminum copper, with a grain size in the half
micron range, titanium, with a grain size in the 0.1 micron range, or
titanium nitride, with a grain size in the 0.05 micron range. The
images we use are from either AFM, immersion lens field emission SEM, or
TEM. We would like the software to be able to reconstruct the grain
boundary network in a fairly reliable manner. However, everything I
have seen so far has been disappointing due to the fact that the grain
boundary network in these images, while easily discerned by the human
eye-brain system, is not well discerned by most software unless the
contrast and S/N ratio exceed what can be obtained in real-world images.
Does anybody out there have knowledge of grain size software that they
are enthusiastic about? MAC or Windows, it doesn't matter to me."
And I am finally getting around to summarizing some of the responses. One individual suggested "Compix", but did not provide a locator for Compix. That person did provide a locator for herself, and that is "mager@unixg.ubc.com"
One individual suggested "Differential Hysteresis" for detecting grain boundaries. This is embodied in the "Pixision" product marketed by JEOL. That person is at "Peters@bsac.uchc.edu"
One individual suggested "PrismView" written by John Russ. That person is at "l_thomas@ccm.pnl.gov".
One individual suggested "Image Tool", and gave the address of the author, "dove@uthscsa.edu".
One individual suggested suggested "Image Pro Plus" and gave the number of Ira Harron at 301-495-3305 X209.
One individual suggested "Optimas". They are at "Optimas@aol.com".
However, the preponderance of opinions was that nothing works like the "eye-brain" system (as opposed the the "hair-brain system) and people actually doing serious image analysis on poorly defined grains in practice outline or trace the grains by hand in order to get meaningfull results. This is such a quagmire that I have hired a summer intern to spend three months full time investigating the subject. I will let you know what we discover.
--John Mardinly
John_Mardinly@ccm.sc.intel.com