Granulometric Filtering

Granulometric Filtering

Author: Dimiter Prodanov

History: version 1.5 11 Nov 2007
- capabilty to launch as a standalone application
- execution time calculation
- brightness correction
- change in distance calculation

1.4 8 July 2007
- suppot of closing operation
- support of Rois
- improved distance calculation

1.3 / 26 March 2005
1.2 / 19 Jan 2004

Source: Contained in the JAR file. To open a JAR file, change the extension from ".jar" to ".zip" and double click on it.

Installation: Download Gran_Filter.jar to the plugins folder, or subfolder, restart ImageJ, then run the plugin using the Plugins/Morphology/Gran filter command.

Description: This plugin performs granulometric filtering of digital images. The algorithm is described in Prodanov et al. 2006 J Neurosci Methods. 15;151(2):168-77.
doi:10.1016/j.jneumeth.2005.07.011
Perhaps the oldest and most frequently used technique in the empirical sciences to quantify the size of solid particles is to use a series of sieves with increasing mesh openings. To quantify the properties of discrete sets of objects Matheron theorized empirical sieving into the formal concept of mathematical granulometry (Matheron, 1975). Granulometry was later applied in image analysis to both binary and continuous tone images (Serra, 1982). In a way similar to sieving grains, pixels comprising an image are "sieved" according to their connectivity to similar pixels imposed by a certain primitive geometric body termed Structuring Element (SE). An integral characteristic of granulometry is the distribution of pixels with respect to the diameters of the used SE-s. A local maximum in its normalized first derivative, the granulometric size density (G(d)), indicates the presence of a number of objects matching the particular SE. Moreover, granulometry can act as a band-pass filter capable of discriminating grains of a certain size based on their similarity to a SE. By exploiting this property, images can be simplified substantially to successfully isolate various classes of objects.
Original image:

Granulometric filtereing with kernels 4 -12

Difference image:

The example is published in Prodanov and Feirabend. 2008 Brain Res. 1233:35-50.
doi:10.1016/j.brainres.2008.07.049
About the user interface:

Type of SE: can be one of circle, diamond, square, hor. line or vert. line
Radius of SE: the radius or half/length [square, lines] of SE
Step: the increse from the Radius - Radius2=Radius+step
Euclicean distance: check to compute the Euclidean distance/area

|Home |Plugins | ImageJ |

Author:	Dimiter Prodanov
History:	version 1.5 11 Nov 2007 - capabilty to launch as a standalone application - execution time calculation - brightness correction - change in distance calculation 1.4 8 July 2007 - suppot of closing operation - support of Rois - improved distance calculation 1.3 / 26 March 2005 1.2 / 19 Jan 2004
Source:	Contained in the JAR file. To open a JAR file, change the extension from ".jar" to ".zip" and double click on it.
Installation:	Download Gran_Filter.jar to the plugins folder, or subfolder, restart ImageJ, then run the plugin using the Plugins/Morphology/Gran filter command.
Description:	This plugin performs granulometric filtering of digital images. The algorithm is described in Prodanov et al. 2006 J Neurosci Methods. 15;151(2):168-77. doi:10.1016/j.jneumeth.2005.07.011 Perhaps the oldest and most frequently used technique in the empirical sciences to quantify the size of solid particles is to use a series of sieves with increasing mesh openings. To quantify the properties of discrete sets of objects Matheron theorized empirical sieving into the formal concept of mathematical granulometry (Matheron, 1975). Granulometry was later applied in image analysis to both binary and continuous tone images (Serra, 1982). In a way similar to sieving grains, pixels comprising an image are "sieved" according to their connectivity to similar pixels imposed by a certain primitive geometric body termed Structuring Element (SE). An integral characteristic of granulometry is the distribution of pixels with respect to the diameters of the used SE-s. A local maximum in its normalized first derivative, the granulometric size density (G(d)), indicates the presence of a number of objects matching the particular SE. Moreover, granulometry can act as a band-pass filter capable of discriminating grains of a certain size based on their similarity to a SE. By exploiting this property, images can be simplified substantially to successfully isolate various classes of objects. Original image: Granulometric filtereing with kernels 4 -12 Difference image: The example is published in Prodanov and Feirabend. 2008 Brain Res. 1233:35-50. doi:10.1016/j.brainres.2008.07.049 About the user interface: Type of SE: can be one of circle, diamond, square, hor. line or vert. line Radius of SE: the radius or half/length [square, lines] of SE Step: the increse from the Radius - Radius2=Radius+step Euclicean distance: check to compute the Euclidean distance/area