Segmentation analysis

Main Menu: Process->Segmentation...

Segmentation is a special kind of ROI where the region of interest is generated based on pixel thresholding. The resulting ROI is a scatter ROI (see “ Region of interest (ROI) Analysis”) which can be saved and reloaded like any other ROI.

Note that segmentation can be performed globally on a whole image or regionally within a ROI. To limit the segmentation to a ROI, draw a freehand or polygon ROI on the image prior to activating the Segmentation function. The ROI can be modified during the segmentation process (i.e. by moving one of the anchors in the polygon ROI) and the segmentation will be updated accordingly.

Example of applying segmentation

Example of applying segmentation to a defined ROI (left) and globally (right). The ROI limited segmentation is automatically performed if a polygon or freehand ROI is active when the Segmentation function is activated.

On activating the Segmentation function, a dialog will appear:

The Segmentation Dialog

The Segmentation dialog window enables dynamic modification of the pixel thresholding values as well as other relevant ROI related options. The red bar indicates the pixel level which is currently segmented (note the segmentation color is always red). The threshold can be adjusted interactively by right-clicking one of the edges of the bar and dragging it to the left/right to expand/reduce the length of the bar. The length of the color bar then reflects the range of segmented pixels which is also reflected in the editable number boxes at either end of the threshold bar. The upper/lower segmentation values can be edited manually by entering the appropriate values in the boxes. The numbers above the boxes indicate the maximum and minimum pixel value in the image.

Max/ Min Segmentation range:

For DICOM images, the segmentation range can be set to include the entire pixel range in the image or just to the windowed range (set with the Image Intensity/Contrast function). In some images, the pixel data may contain some outliers which will tend to make the pixel range of interest limited to a small part of the dynamic range. In this case, set the Window range to the range of interest and select Window range rather than Pixel Range under this option. For most images, however, this setting will have little effect on the segmentation.

Update segmentation on image scroll:

When this option is checked, the segmentation will be updated according to the pixel max/min in the new image when a new image is made current by image scrolling. If this option is unchecked, the pixel range will be kept unchanged on image scrolling.

Segmentation stats:

This window displays statistical data about the segmented region (similar to the statistics shown in the ROI Statistics window). Pixel mean, SD, segmented area and number of segmented pixels is shown. Pixel area is only correct for DICOM images where the pixel size is available in the DICOM header.

Enable pixel masking:

The selected region will be masked out and optionally set to a fixed value specified pixel masking value. Alternately, the non-masked region is set to the specified value. The masking operation can be applied only to the current slice or all slices, depending on the option ‘Update segmentation on image scroll’.

Save ROI to Buffer:

Save the segmented pixels as a scatter ROI to the ROI Buffer (see “ The ROI Buffer ”). The segmented ROI can then be treated like any other ROI in the buffer.

Save ROI->Buffer scope:

Current slice:

The segmentation setting will only be applied to the current slice.

All slices:

All slices in the series will sequentially be scrolled through and the segmentation will be applied to all slices in turn and saved to the ROI Buffer. This is an efficient way of applying the same segmentation setting to many slices and save the resulting segmented pixels to the ROI Buffer.

Save segm:

The segmented pixels are save as a single ROI file (with extension .seg). The ROI can then be reloaded as a scatter ROI on any dataset. This is a very useful function for e.g. creating ROI s from pixel segmentation which is subsequently applied in e.g. dynamic ROI analysis. Once the segmentation is reloaded as a scatter ROI, all ROI functions can be accessed like for any other ROI. Image size and zoom/pan settings applied when the segment is saved will be recorded in the .seg file and the same settings will be applied upon loading the .seg file.

Reset:

Resets the segmentation range to default values.

Close:

Close the Segmentation window.

Example of Segmentation limited by a Polygon ROI.

Example of Segmentation limited by a Polygon ROI. The segmented pixels are shown in red, and the segmentation range is adjusted interactively in the ROI from Grayscale Segmentation window (botton left). The ROI Statistics window (bottom right) displays the ROI statistics for the polygon ROI defining the boundaries of the segmentation process. The upper right window is the ROI Buffer window showing the results of saving the segmentation as a scatter ROI to the ROI Buffer using the Save ROI to Buffer function in the segmentation dialog.

Resulting image obtained after saving the segmentation

Resulting image obtained after saving the segmentation in “ Example of Segmentation limited by a Polygon ROI. ” as a scatter ROI and reloading the ROI (as a .seg file) to the same dataset. The segmentation is now available as a scatter ROI, and all ROI functionality is available. Note that if the original image was zoomed or enlarged when the segmentation was performed, the same zoom - and enlargement factors will be applied when loading the ROI buffer. Also note that each point in the scatter ROI is displayed as 2 by 2 red pixels, while in the original image, each point in the segmentation was displayed as a single red pixel.

Segmentation analysis

Related topics: