Histographic Segmentation

1. Choose Tools > Histographic Segmentation on the menu bar.

   The Histographic Segmentation dialog appears.

2. Choose the dataset that you want to plot on the X axis in the X axis dataset drop-down menu.
3. Choose the dataset that you want to plot on the Y axis in the Y axis Y dataset drop-down menu.

   

   NOTE The 2D and 3D implementation of Sobel of the selected dataset is available by default for plotting on the Y axis.

4. Choose a mask in the Mask drop-down menu, if available and if required. If a mask is selected, only data points within the mask will be plotted.

   

   NOTE Click the Refresh button to update the list of available datasets and masks, if you added new data to the Data Properties and Settings panel while you were working in the Histographic Segmentation context.

5. Select the required number of bins — 64, 128, 256, or 512 — for the histogram, if required.

   

6. Click the Compute Histogram button. The datasets are plotted in the dialog.
7. Adjust the display of the histogram, if required.

   You can change the color map applied to the histogram, as well as display contour lines and adjust the opacity of the ROIs (see Modifying Histograms).

8. Create the required number of regions of interest (see Creating Regions of Interest).

1. Click the Options button in the Histographic Segmentation dialog.

   The Histogram Options dialog appears.

2. Select the required options, as described in the above table.
3. Do one of the following:
   • Click Apply to preview your changes.
   • Click OK to apply your changes and close the dialog.
   • Click Cancel to discard all changes and close the dialog.

The plot tools available in the Histographic Segmentation dialog let you pan and zoom the figure, as well as edit the titles and attributes for each axis. You can also save the figure to a number of different file formats.

Plotting tools

	Icon	Description
Reset		Resets the original view of the figure.
Pan		Pans or zooms the figure as described below. Pan… Click with the left mouse button and then drag to pan the figure. Zoom on the Y axis… Click with the right mouse button and then drag up and down to Zoom in and Zoom out on the Y axis. Zoom on the X axis… Click with the right mouse button and then drag left and right to Zoom in and Zoom out on the X axis.
Zoom		Zooms to a selected rectangle, which can be defined by dragging your cursor over the area that you want to zoom.
Save		Saves the figure as a bitmap image, vector graphic, or in the PDF file format. The figure can also be saved as raw data or PGF code for LaTeX. Standard image files (*.jpeg, *.jpg, *.png, *.tif, *.tiff extensions)… Saves the figure as a bitmap image in the screen resolution. Postscript files (*.eps, *.ps extensions)… Saves the figure as an encapsulated postscript or postscript file. These types of files have a selectable resolution and provide high-quality graphics for publications. PGF code for LaTeX (.pgf extension)… Saves the figure in the Portable Graphics file format. Raw RGBA bitmap (*.raw, *.rgba extensions)… Saves the figure as a raw bitmap image file, in which the file contains only a list of pixel colors and nothing else. Scalable vector graphics (*.svg, *.svgz extensions)… Saves the figure in an XML-based vector image format. The SVG specification is an open standard developed by the World Wide Web Consortium (W3C). SVG images and their behaviors are defined in XML text files. Portable document format (*.pdf extension)… Saves the figure in the Adobe PDF file format.

1. Click the Create New ROI button in the Regions of interest box.
2. Add data points to the selected region of interest with the Brush tool.

   You can choose to work in the Histogram or Image domain or both (see Working in the Histogram or Image Domain).

3. Add the required data points to the selected region of interest as follows:
   • Use the scroll wheel on your mouse to change the brush size, if required.
   • Hold-down Left Ctrl (or your configured key) to add data points to the selected region of interest.
   • Hold down Left Shift (or your configured key) to remove data points from the selected region of interest.

     NOTE You can change slices while you work. Click the Refresh Preview button in the Image domain box to update the view of a new slice.

4. Create additional regions of interest, if required.
5. Export and and/or expand the regions of interest that you created (see Exporting Regions of Interest and Expanding Regions of Interest).

Options for working in the Histogram domain are available in the Histogram domain box, shown below.

Histogram domain box

Histogram domain options

	Icon	Description
Undo		Undoes the last operation that you performed in the Histogram or Image domain. NOTE Some operations or actions cannot be undone, such as reinstating a removed region of interest or exporting an ROI. If you can't undo an operation, the Undo button will remain inactive.
Redo		Redoes the last operation that was undone.
Brush		The Brush tool lets you make selections on the histogram and is dependent on the Use range setting. Use the scroll wheel on your mouse to change the brush size. Hold-down Left Ctrl (or your configured key) to add data points to the selected region of interest. Hold down Left Shift (or your configured key) to remove data points from the selected region of interest. NOTE Whenever you draw in the histogram, the results are updated in the previews automatically.
Automatic clustering		Creates the number of specified regions of interest through a K-Mean or Mean Shift clustering algorithm. You should note that the Use range option can be selected for automatic clustering. Results of automatic clustering [1] http://scikit-learn.org/stable/modules/clustering.html#k-means [2] http://scikit-learn.org/stable/modules/clustering.html#mean-shift
Use range		You can use this feature to define a domain of intensity, which is highlighted a constant color. Any voxel with a value outside the defined range will not be selectable with any of the Brush tools. Drag the left or right Range sliders to change the minimum and maximum values of the intensity range. Drag from the middle of the Range slider to translate the intensity range values.
Dilate		Expands the region of interest.
Erode		Shrinks the region of interest

Options for working in the Image domain are available in the Image domain box, shown below.

Image domain box

Image domain options

	Icon	Description
Brush		The Brush tool lets you make selections on a preview image and is dependent on the Use range setting. Use the scroll wheel on your mouse to change the brush size. Hold-down Left Ctrl (or your configured key) to add data points to the selected region of interest. Hold down Left Shift (or your configured key) to remove data points from the selected region of interest. NOTE You can change slices while you work. Click the Refresh Preview button to update the view of a new slice.
Refresh Preview		Updates the view of a new slice in the preview window.
Expand ROIs		Creates a series of ROIs that fully segment the selected dataset (see Expanding Regions of Interest).

1. Check your segmentation on representative slices within the dataset with which you are working and then make any necessary adjustments, recommended.

   After you scroll through the dataset, you must click the Refresh Preview button to update the preview.

2. Click the Export Selected ROI button.
3. Enter a name for the selected data in the Export ROIs dialog, optional.

   If a name is not entered, then the names assigned in the Histographic Segmentation dialog will be retained with the designation Hist. Seg. appended at the beginning.

   If a name is entered, the new name plus the names assigned in the Histographic Segmentation dialog will be added to the new data.

4. Select the required export options as indicated below.
   • Choose Export selected ROIs one-by-one to export each selected region of interest as a new ROI.
   • Choose Export selected ROIs as a multi-ROI to export all selected regions of interest as a single multi-ROI.
5. Click the OK button.

   The exported region(s) of interest appear in the Data Properties and Settings panel (see ROI Properties and Settings and Multi-ROI Properties and Settings for information about working with regions of interest and multi-ROIs).

The watershed algorithm, whose name comes from an analogy with hydrology. has many applications in image processing, such as automated object segmentation or separation. Starting from user-defined markers, the watershed algorithm treats voxel values as a local topography (elevation). The algorithm then floods basins from the markers, until basins attributed to different markers meet on watershed lines that correspond to the desired segmentation.

This algorithm depends on the following inputs:

1. Labeled regions that are used as seed areas for the flooding. At the end of the process there will be as many separated objects as markers labeled differently.
2. An image that plays the role of the landscape height field or altitude map that controls the flood progression and finally the location of watershed separations.
3. A distance map that provides the priority map input for a watershed process.

Different applications can be achieved by carefully choosing the inputs.

Watershed options

	Description
Landscape	The dataset that will provide the local topography for the watershed algorithm.
Use mask	If selected, the mask used to compute the histogram will be applied to the watershed calculation.

1. Add the required regions of interest that define the inputs used for the expansion.

   NOTE You must create a least two regions of interest to use the watershed expansion scheme.

2. Check your segmentation on representative slices within the dataset and make any necessary adjustments, recommended.

   After you scroll through the dataset, you must click the Refresh Preview button to update the preview.

3. Select the required algorithm in the Expand ROIs drop-down list and then click the Start button.
4. Select the dataset that will provide the landscape for the segmentation in the Watershed dialog. The Sobel of the selected dataset is recommended, if available.

   

5. Select the Use mask option, if you want to limit the segmentation to only data points within the previously selected mask.
6. Select the required geometry — 6, 18, or 26 — in the Geometry box.

7. Click the OK button.

   New regions of interest generated from the seed points are added to the Data Properties and Settings panel (see ROI Properties and Settings).