These modes are used to segment the images and create the TAG data. Each segmentation technique has its advantages and its inconvenients.  When one of these modes is active the 3D tools will not be available. Any window associated with a TAG Mode will be in 2D mode (either Mode ONE or Mode ALL). The available modes are:

 

 

•Edit

This is a simple "paint box" that enables to edit the pixels of the TAG images. It is useful to re-touch any small errors that can be left after having used the other segmentation modes.

 

•Geometrical Masks

The geometrical masks are useful as post processing.  You can use the masks to increase the number of tags present in an image.  If, for example, you want to differentiate between left and right organs, you can segment the organs using the same TAG values for left and right  and after that create a rectangle mask over the right part of the body and "add" the TAG value associate with the mask to the TAG image to create a different TAG value for the right hand side organs.

 

•Morpho

The Mathematical Morphology mode is very useful for segmenting some type of images.  Any tissue that has a big enough surface and a well defined gradient will be easy to segment with this Module.  The main advantage of Morpho is that the segmentation technique is based on the variation of the pixel values (the gradients), not the values themselves.  So this makes it ideal for MR images where the pixel values are not uniform over the image.  A good example would be sub-cutaneous fat in MR images.  It is not very good for segmenting very small regions however, so it would not work well with the visceral fat. If you work with a 3D dataset, Morpho propagates very well from one frame to the next.

 

•Region Growing

The Region Growing Mode is based on thresholding.  So it works directly with the pixel values.   It has 3 sub-modes: "Paint", "Grow 2D" and "Grow 3D".  "Paint" is very useful when you want to threshold only a small region in the image. For example, it is the ideal tool to segment visceral fat. Just select a big brush, fix a threshold that will capture the fat (you can adjust threshold values with the mouse wheel) and "paint" your fat in the abdominal cavity with the brush.  "Grow 2D" grow a region starting from the cursor position. All adjacent pixels that fall within the threshold range, and respect the brush constraint, are tagged. "Grow 3D" is similar to the 2D version, but also propagate through all the frames that belong to the same parent in the database tree. Unfortunately, even though the option is present, the Region Growing mode does not propagate well from one frame to the next.

 

•Snakes

Snakes, also called "Active Contours", work well with tissues that have a smooth well defined contour such as bones in CT.  It is also the ideal mode to compute the circumference of the body.  Just place a few points outside the body and minimize the curve.  You can then transform the snake in a geometry and get the length of the curve from the 3D measurement tool. If you work with a 3D dataset, the Snakes propagate very well from one frame to the next.

 

•Thresholding

The Thresholding Mode applies a global threshold to the complete image.  Apart from a few cases in CT images, global thresholding is not the ideal segmentation tool.  Since it is global, it tends to tag tissues that should not have been tagged if you work with images containing multiple organs.  For simple images, such as CT of the members, it can be very useful.  The threshold are based directly on the pixel values, so in CT you can fix your threshold directly on the Houdsfield values. However, for more complex images with multiple organs, you have more control with the Region Growing tool.

 

 

From the Keyboard

 

There are a few keyboard keys that have been mapped to commands that can be used in all the TAG modules.

 

 

 

Technical Note:

 

 

From the Command Line

 

A number of commands affect the tags. These can be found in:

"SliceO_System\Brush.dll" section of Appendix B.

"SliceO_System\Tag.dll" section of Appendix B.

"SliceO_System\Undo.dll" section of Appendix B.