-
Notifications
You must be signed in to change notification settings - Fork 16
MRI_Count_Spot_Populations_Tool
The tool detects and and counts the spots (or blobs) in an image. It has been created for the counting of bacteria colonies in in Petri-dishes. It separates the spots into two populations and counts each population individually. The populations are separated by the area of the spots. The tool uses expectation maximisation clustering [1] from the weka software [2].
|
|
|
|
To install the tool save the files Count_Spot_Populations_Tool.ijm and results_table-clusterer.py into the folder macros/toolsets of your FIJI installation.
Select the "3D_nuclei_clustering" toolset from the >> button of the ImageJ launcher.
- the first button opens this help page
- the c-button starts counting and clustering of the spots on the active image
- the d-button runs a DoG-filter on the active image
- the r-button runs the clustering by area on a results table
- the p-button plots a histogram and the distributions of the spot areas
- the e-button opens an extras menu, from which example images can be downloaded
Open an input image. The tool will detect dark spots on a bright background. If necessary invert the contrast of the image (Edit>Invert) before running the tool. Set the options for the detection of the Petri-dish and the spots (see below). Make sure that Area
is selected in Analyze>Set Measurements...
. Press the c-button to run the analysis.
The clustering can also be run on a results table, using the r-button. The results table must contain the column Area
. The plotting of the area-histogram and distributions with the p-button is also possible in this case.
Right-click on the c-button to open the options dialog.
- Petri-dish margin
- Allows to exclude a smaller or bigger part that corresponds to the border of the Petri-dish. Set to zero to use the whole Petri-dish or the whole image
- threshold method
- The auto-thresholding method used for the spot-segmentation.
- min. circularity
- Allows to filter spots by their circularity.
- fit ellipse
- If selected an ellipse will be fit to each segmented spots.
- contrast
- Select the contrast of the image: auto, normal or inverted.
- min. diameter of spots
- The minimum diameter of the spots
- max. diameter of spots
- The maximum diameter of the spots
- color cluster one
- the color in which the distribution of the smaller areas is displayed
- color cluster two
- the color in which the distribution of the bigger areas is displayed
- color histogram
- the color in which the histogram of the areas is displayed
- histogram bin width
- the bin width of the area-histogram
- distribution line width
- the line width for the distributions in the plot
-
Yong Gyu Jung, Min Soo Kang & Jun Heo (2014) Clustering performance comparison using K-means and expectation maximization algorithms, Biotechnology & Biotechnological Equipment, 28:sup1, S44-S48, DOI: 10.1080/13102818.2014.949045
-
Eibe Frank, Mark A. Hall, and Ian H. Witten (2016). The WEKA Workbench. Online Appendix for "Data Mining: Practical Machine Learning Tools and Techniques", Morgan Kaufmann, Fourth Edition, 2016.