FeatureJ: Options

General Description

This plugin enables setting global parameters used by the different plugins in the FeatureJ package.

Description of Dialog Components

Separate smoothing and differentiation. Except for statistics, all features computed by the FeatureJ plugins are based on image derivatives. To allow for feature analysis at different scales, these derivatives are by default computed by means of convolution with derivatives of the one-parameter Gaussian kernel. It is known that when increasing the scale, this is, in the limit, theoretically equivalent to first convolving the image with the Gaussian kernel and then taking finite differences. When this option is selected, derivatives are computed using the latter approach. It greatly reduces computation time, as the convolution operation is carried out only once. However, it is theoretically less accurate at lower scales and numerically inaccurate at higher scales.

Isotropic Gaussian image smoothing. By default, the FeatureJ plugins assume that image elements (pixels/voxels) are of size 1.0 in all dimensions. When this option is selected, the plugins use the pixel width/height/depth as specified in ImageJ > Image > Properties. Specifically this means that each dimension is assigned its own Gaussian (derivative) smoothing scale, computed by dividing the scale as specified in any of the plugin dialogs, by the specified pixel/voxel size for that dimension.

Close input images after processing. If selected, ImageJ always closes the input image processed by a FeatureJ plugin, after the plugin has finished. This is useful when you need only the result images and not the original input images. Since the Statistics plugin does not produce result images, this option does not apply to that plugin.

Save result images before closing. If selected, ImageJ always asks whether or not you want to save result images before they are closed. This is useful when you want to exclude the possibility of accidentally closing an image.

Log messaging. If selected, the FeatureJ plugins will generate log messages, which appear in the log window of ImageJ. These messages give you more details on what is going on behind the scene.

Notes regarding log messages: most of them originate from the ImageScience library. In this library, the "c-dimension" (channel index) is called the "s-dimension" (spectral parameter). Also, indexing of image elements always starts at 0, as opposed to ImageJ and Image5D, which start at 1 in the z-, t-, and c-dimension. Thus, in the log window, these indices are reported shifted by 1.

Upon pressing "OK", the options are stored by ImageJ and saved to its preferences file when quitting. Every time a FeatureJ plugin is used, these preferences are used. This facilitates adapting the behavior of the package to your wishes.