next up previous contents
Next: Texture Recognition Up: MR Image Segmentation Previous: Introduction   Contents

Practical

The software you should use for this exercise is in 7.NMRSEG.

The image calculator contains two buttons, scat and iscat. The first of these produces a 2D float scattergram image from the selected region of interest (ROI) of a complex image, where the x axis is the real pixel value (0-256) and the y axis is the imaginary pixel value (0-256).

The iscat button takes a float scattergram and a complex image and generates a new float image by replacing the image pixels with the corresponding value within an identified region of interest (ROI) of the scattergram. Using these two buttons it is possible to perform a dual valued (complex) image segmentation.

Read in the dual_echo1 and dual_echo2 images from the images directory and form a composite complex image in the image calculator (remember to normalise the image to occupy the available grey level range used in the scattergram). This can be as simple as using the two images as the real or imaginary components directly, but you may get better results from constructing alternatives such as sum or difference images. Produce a scattergram of the complex image for regions containing

Try to get a feel for where the various tissue types appear in the scattergram.

Now, using the iscat process as a starting point, attempt to produce three segmentations of the image for the various identified tissue types.

Describe the following in your report:


next up previous contents
Next: Texture Recognition Up: MR Image Segmentation Previous: Introduction   Contents
root 2017-11-20