Set Up a Scan
Click a purple button on the FracLac panel.
- The top button brings up an option panel
for scanning images or ROIs for the box counting
dimension (DB), lacunarity, and measures
based on the convex hull enclosing the pixelated part
of an image.
- Select the purple Sub Scans button to find
local dimensions over parts of an image.
- Select an ANALYZER scan
to automatically detect particles using
ImageJ's ParticleAnalyzer.
- Select a RECTANGULAR/BLOCK
scan to use non-overlapping blocks.
- The SLIDING BOX LACUNARITY scan button scans using
an overlapping (potentially exhaustive) scan
which measures heterogeneity (as opposed to the
nonoverlapping scan of regular box counting).
- The MULTIFRACTAL ANALYSIS button
delivers the generalized dimension spectrum (DQ
and ƒ(α)).
Scan an Image
Click a blue button on the FracLac panel.
- After the scanning options are set for the
first time, the currently selected scan type
is displayed in red on the
bottom of the FracLac panel and blue action buttons appear.
- "Scan Current Image or Roi" analyzes the currently
displayed image or roi selected on it.
- The "Select and Scan Files" button
analyzes images without opening them,
using the currently selected type of scan and settings.
Use this button to select multiple files at once.
- The third blue button analyzes ROIs that
have previously been stored in the Roi Manager
(Analyze>Tools>Roi Manager).
View and Save the Results
The results are displayed in the Results
Table and in graphics, depending on the settings from Step 1.
Several versions of the fractal dimension are provided.
The columns listing the average and
minimum cover DBs
reflect the results for multiple origins. The smoothed
DB is found from smoothing horizontal
intervals from the regression line by removing
successive box sizes yielding the same number of boxes.
See the User's Guide for more information.
NOTE:
- Options include type of box size
series
as well as other scanning features that can affect the results
for different types of images. For instance, a Relative
Series is calculated relative to the largest box using exact factors; thus,
setting the maximum box size to 48 generates the box sizes
{1,2,3,4,6,8,12,16,24,48}, whereas setting it to
49 generates {1,7,49}. Refer to the current User's Guide
(2006) for more information.
- For multifractal analyses and sliding box lacunarity analyses,
certain settings, in particular the minimum resolvable box size,
may need special attention.
- If the red and green parts of the MF graph of ƒ(α)
do not meet in a continuous curve or if they appear to
cross over, for instance, sampling may be inappropriate.
To correct this, try using a scaled series,
increasing the
minimum grid size in pixels, or increasing the
number of box sizes.
- For sliding box lacunarity,
it is important in some cases to keep
the minimum resolvable size (i.e., minimum grid size in
pixels)high enough (e.g.,usually greater than 3).
In addition, using a nonexhaustive scan often yields the same
basic results as an exhaustive scan for preliminary quick
analyses. To use a nonexhaustive scan, keep the slide
factors greater than one (e.g., 5); setting the x and y
slide factors to 1 does an exhaustive scan.
- Note that the number of box sizes can affect the result,
especially in a multifractal analysis.
To reduce processing time in both multifractal and lacunarity
analyses, it is often helpful to use a relatively small number of
box sizes. Do this by setting the number
of different box sizes to 10 or 20,
for instance.
- For a random mass multifractal analysis, you may need to
adjust the maximum percent to 100%, ensure the number of box
sizes is not too high and not too low, and ensure the
minimum grid size in pixels is high enough (e.g., >5
or 10).
References for FracLac
-
2005 - “Fractal Analysis of Microglial Morphology”, Karperien, Lucas,
Aurel, and Jelinek in Bourgine, Kepes, and Schoenauer (eds.), poster,
Proceedings of the European Complex Systems Society – Towards A Science
Of Complex Systems.
- 2004 - “Fractal Analysis Quantitates Overt and Subtle Effects of
Naloxone and Lipopolysaccharide on Cultured Rat Microglia”, Karperien,
Jelinek, Bossomaier, in Proceedings of Complex 2004, the 7th Asia-Pacific
Conference on Complex Systems, Cairns, Australia.
- 2004 - “Differentiating Grades of Microglia Activation with Fractal
Analysis”, Jelinek, Karperien, Bossomaier, and Buchan, in Proceedings of
Complex 2004, the 7th Asia-Pacific Conference on Complex Systems,
Cairns, Australia.
- 2002 - "MicroMod - an L-systems Approach to Neuron Modelling".
Jelinek, Karperien, Cornforth, Cesar, and Leandro, in Sarker, McKay, Gen and
Namatame (eds.), Proceedings of the Sixth Australia-Japan Joint
Workshop on Intelligent and Evolutionary Systems, Canberra, Australia.