This dialog is activated via the Setup button in the Shape Measurements pane or from the Output tab on the Particle & Pore Analysis pane. Both places there are shortcuts to go directly to the desired tab in the setup dialog. The dialog consists of three tabs: Parameters, Statistics and Charts.
On each tab the items are organized in a grouped tree structure (for Statistics there is only one level).
A description of all measure shape parameters is found in the Reference Guide of this manual.
The Parameter groups are ordered so that all Parameters within one group have the same dimensionality. To toggle any item on or off fast click the corresponding check box (clicking slowly will not work!). All items at one branch level can be enabled or disabled by toggling their header item. The same counts for unit selection, which can be done per Parameter, for a group of Parameters or for all Parameters at once.
In addition to simply enabling or disabling each statistical quantity in the same way as for Parameters, it is possible to specify the level for the Percentile (=percentile x 100).
Charts – Histograms and Scatter Plots - are mostly relevant when there are many Measure Shapes, e.g. when they have been generated by the Particle & Pore Analysis module. However, they can also be used for manually created measure Shapes. The Charts are enabled or disabled in the same way as Parameters, see above. To have display any selected Charts click Show Charts in the menu drop down on the Setup button on the Shape Measurements pane toolbar or the Charts button on the Output tab of the Particle & Pore Analysis pane.
The histogram shows the statistical distribution of a given Parameter. In an ordinary Histogram all values of the Parameter of interest are first sorted. Thereafter the values are grouped in intervals - bins - of equal size. The histogram is a graph with columns representing the bins (intervals). The height of each column is proportional to the number of values found in the corresponding bin (interval).
The y unit of the histogram can be either percentage (%), fraction, or count. Count is simply the actual number of Measure Shapes for which the Parameter of interest has a value in the interval associated with each bin. If all columns were stacked they would sum up to the total number of Shapes (Shape Count). Fraction is the number relative to the total number of Shapes. In this unit the sum of all fraction columns adds up to unity. Percent (%) is simply fraction times 100, i.e. the sum of all columns adds up to 100%. The y axis unit is set in the View Settings for Histogram Windows (right click on the Histogram to get there).
By default Histograms will have 20 bins distributed over an interval given by the Minimum and Maximum value of the Parameter. The user can define the minimum and the maximum value and the number of bins in the histogram. This is done by clicking the three small dots next to the check mark. This brings up a small dialog where these values can be set.
In Area and Volume Weighted Histograms instead of counting how many Shapes fall into each bin for the Parameter of interest, the Area or Volume, respectively, of the Shapes which fall into each bin is summed. This means that in the Area Weighted Histogram if all columns were stacked they would sum up to the Total Area covered by the Shapes. In the Volume Weighted Histogram if all columns were stacked they would sum up to the Total Volume of all the Shapes.
Example of how many small Shapes are dominating the ordinary (counting) Histogram and how they are suppressed in the Area Weighted Histogram.
Weighted histograms are useful for illustrating “what is important”. As a thought example let us consider an image of a filtration membrane. Let us say that we find 1000 pores using the Particle & Pore Analysis module and that (very realistically) 95% are below 50 nm in diameter, and the remaining 5% all have diameters above 500 nm. The total area covered by the small pores will then be less than 1% of the total pore area! Still in the ordinary histogram the small pores will dominate with high columns in the low end of the Histogram. But those pores which really matter for the flow through the membrane, namely the big ones can hardly be seen in the histogram! In the Area Weighted Histogram, as the name suggest, weight is given to those Shapes which have the larger Area. The Area or Volume Weighted Means and Standard Deviations reported as statistical quantities in the Shape Measurements pane provide numbers describing the weighted distributions.
SPIPTM also provides scatter plots (or diagrams) where one parameter can be plotted against another parameter from the Shape Measurements result section in a graph. In this way correlations can easily be visualized.
Further, the scatter plots provide a tool for cluster analysis. In cluster analysis one parameter is plotted against another parameter. If the points in the scatter diagram tend to cluster in separated dense clouds it indicates that the image features associated with the Shapes can be categorized according to those two parameters.
Example: Maximum Height vs. Diameter. Most Shapes clutter along a line in the left side of the graph. The linear dependence indicates that the underlying features are three dimensional, i.e. the height depends on the lateral dimensions. The Shapes scattered in the middle and the right side of the graph must be of different nature.
The parameter plotted on the x-axis of the Scatter Plot is the one listed in the Charts tab in the Parameters/Statistics/Charts setup dialog, for which the Scatter Plots check box is marked. The second parameter, plotted along the y-axis, is selected by clicking the three dots to the right of the check mark. Then a new dialog with yet another tree list of the parameters pops up. Select the parameter by clicking in the Enabled column.
Learn more about classification in the Shape Classification section.