Local thresholding of regions.

This plug-in computes a threshold for each region using a given method.

The only available method is

First the radial distribution of the object is computed (see plugin 3D Radial Distribution), as concentric circles centred on the seed define growing regions of interest in which mean intensity values are measured. A Gaussian fit of the radial distribution is computed in a given radius around the seed (Radius Max, in pixels)). The standard deviation of the fitted Gaussian curve is used to define the threshold. The user enters a factor which is applied to the standard deviation to define the value of the threshold (standard Deviation value). As a rule of the thumb, a factor 1.17 will bring the threshold to the full width at half maximum while factors 2 and 3 will fill about 90% and 99% of the curve surface, respectively.

The parameter

This procedure is adapted from the method Gaussian Fit of the algorithm 3D Spots Segmentation.

This plug-in computes a threshold for each region using a given method.

The only available method is

*Gaussian Fit*of radial distribution of intensities:First the radial distribution of the object is computed (see plugin 3D Radial Distribution), as concentric circles centred on the seed define growing regions of interest in which mean intensity values are measured. A Gaussian fit of the radial distribution is computed in a given radius around the seed (Radius Max, in pixels)). The standard deviation of the fitted Gaussian curve is used to define the threshold. The user enters a factor which is applied to the standard deviation to define the value of the threshold (standard Deviation value). As a rule of the thumb, a factor 1.17 will bring the threshold to the full width at half maximum while factors 2 and 3 will fill about 90% and 99% of the curve surface, respectively.

The parameter

*Local Coefficient*allow to smooth the threshold values among all the spots of the image to limit the effect of the local threshold:- if it is set to 0: the same threshold (mean of all thresholds) will be applied to all spots
- if it is set to 1: each spot will be thresholded with its own local threshold
- between O and 1: a smoothed threshold is computed for each spot: smoothedThreshold = localThreshold * (1 - coefficient) + mean * coefficient

This procedure is adapted from the method Gaussian Fit of the algorithm 3D Spots Segmentation.