Set the threshold for image #1 such that just the bone is highlighted in red (about 136).
Scroll through the data set and note that the threshold chosen, although appropriate near the beginning and end of the series, is too low for those images in the middle of the scan.
To correct for the variation in average gray levels, and thereby establish a constant threshold value for bone across the data set, you will want to use the Gray Level Point Operations Adjust function. Adjust produces an image with specified mean and standard deviation of gray levels. However, you do not want the Adjust operation to be applied to the entire image, just those pixels that define the patients body.
To restrict the Adjust operation you must first create an ROI in each image that segments the body from the background. ROIs in Velocity 2 Pro are implemented using masks. Each gray level image is automatically assigned a file of mask data the first time the software opens an image set. Each mask data file has eight independent, 1- bit masks available, numbered Mask One through Mask Eight. The software lets you define these masks as needed for your various ROI operations. The contents of a mask are independent of the contents of all other masks throughout the data set.
Masks are selected by the buttons in the ROI Mask box labeled One through Eight. You create masks by using one of the Logic tools, such as Flood Fill, Flood Fill All, Draw Rubberband Line, Draw Rubberband Polygon, Freehand Paint or Seed Fill.
Select image #1 and run the Flood Fill function.
Vary the threshold meter and highlight all pixels throughout the body (about 46).
Scroll the filmstrip window and verify that this value is appropriate
for the
entire set.
Now, return to image #1, click on Mask One in the ROI Mask box. It should be highlighted by a green border.
Click on the Clear All button in the Mask Ops box. This clears all data from Mask One of every image of the set.
Even though the mask may not be in use, as indicated by the absence of a check mark on the button, it is good practice to clear data from the mask to be sure it is empty.
In the next few steps, you will create an ROI in Mask One of each image using the Flood Fill function in 3D Fill mode. In this mode, you first set threshold in one image and then, when you click on a red area in that image. The software finds all connecting pixels of the same threshold throughout the entire image set.
Check mark the 3D Fill button. Also, make sure that the check mark is on in the Draw( Erase) button, to enable drawing of pixels in the overlay planes.
Perform the 3D Flood Fill function on image #1 by clicking anywhere in the red pixels in the patients body. After a few seconds the body will be highlighted in the overlay color (default is yellow).
Click on the Set Mask button in Mask Ops. Mask One should now be check marked. To verify that the mask has been set for other images in the set, use the Video Loop controls.
When done, you will have created the ROI needed to run the gray level adjust and filtering operations.
In the next few steps, you will learn to create a script and then use it to apply gray level adjust and median filtering to the image set. Median filtering is useful in removing single pixel noise from the data.
Click on the Gray button in the control panel toolbar.
Make sure that you have image #1 displayed.
Click on the Adjust button.
Click on Mask One in the ROI Mask box to turn the check mark on.
Turn scripting on by clicking on the Record Script button (image display window, top toolbar).
Click on the Perform button in Gray Level Point Operations to apply the Adjust function.
Switch to the filtering tools by clicking on the Filter button in the control panel toolbar.
Click on the Median button.
Select Mask One in ROI Mask.
Click on Perform.
Now, click on the Save Image button (very important).
Having applied the desired steps to the first image, you can apply them to the rest of the images by running the script.
Click on the Run Script button (image display window, top toolbar). Wait for processing to complete.
Return to the Logic tools and run the Flood Fill function.
If an overlay is visible, click on the Erase Overlay button.
Vary the threshold and note the value that gives the best segmentation of the pelvic bones (we recommend 165).
While using the Logic tools, select image #1 and then click on Mask Two.
Click on Clear All to clear all mask data from Mask Two.
Be sure 3D Fill is on (Modifiers).
Use Flood Fill in 3D Fill mode to first create a mask of just the table pad. To do this, vary the threshold until the pad is highlighted in red in all the images of the set (we recommend 87), and then click on the pad in any image (the narrow curved strip at the bottom of each image).
Save the set of masks to Mask Two (Set Mask button).
Turn 3D Fill off.
Return to image #1 and click on the Record Script button.
Since you have previously recorded a script, the program will ask if
you want
to append new commands to the current script. Click on No.
Click on the Erase Overlay button.
Select Mask Two and then click on the Read Mask button. The mask for the table pad should be visible.
Click on the Dilate button four or five times. This expands the mask sufficiently to completely overlap any table pad pixels.
Select Mask Three and click on Clear Mask.
Save the dilated mask to Mask Three by clicking on Set Mask.
Now, run the script to automatically create the dilated mask for the rest of the data set.
The pelvic bones can be reasonably represented by a constant threshold (of about 165), and we have a mask to get rid of the table pad.
| This information is available in alternative formats upon request by individuals with disabilities. Please send email to alt-format@msi.umn.edu or call 612-624-0528. | ||||||||||||
|
HOME |
QUESTIONS |
FEEDBACK
|
||||||||||||
|
||||||||||||