In this step, the mixer class is activated and use to mix the water and fat images using "Fat - Water".
|
|
|
|
|
Mixer channel order need to be changed |
|
Correct mixer equation |
What the script did for you:
The database tree is now composed of "Root", "Mixer" and "Frames".
The "Mixer" interface is open.
The mixer equation is set to "P1 - P2".
The bypass of chanel 1 (Fat) is mapped to the "f" key
The bypass of chanel 2 (Water) is mapped to the "w" key
The "Color Scheme" tool is open.
What to do:
|
If nothing is wrong and the images are displayed correctly, you have no need to do anything at this step. We only open this step to make sure the mixer is properly set.
However, it is possible that the order of the series (Fat and Water) are inversed! Also, if you read an entire study, you probably have a few other series, seen as additional channels in the mixer, that are not really used (In phase, Out of phase, Fat fraction...)
We will fix this using the "order" tab of the Mixer interface.
|
|
|
•In the "Order" tab, make sure the first channel (the top line) is "Fat", the second is "Water" (or sometime "Fat suppressed") and you can close any other channels.
|
|
|
|
|
|
"Order" tab with wrong order and to many channels |
After cleanup and re-order of the Fat/Water channels |
The Interface:
The mixer interface has 3 pages: "Channels", "Order" and "Mixer".
|
|
|
|
The "Channel" page help use organize how the images are mixed together. To mix the images, we need to separate them in groups (the channels) and inside each of these groups, we need to identify slices that are to be associated with one another. The DICOM tag taht will be used to discriminate the channels is different for Ge and Siemens! With Ge images we can use the tag (0008,103E) "Series Description" since the description for each series in the IDEAL sequence is different. For Siemens, we use the tag (xxxx,yyyy) "???". For both Ge and Siemens we can use the same DICOM tag (0020,1041) "Slice Location" to associate slices that belong together. Since we can have multiple channels but we create ony one TAG image, we need to know with what channel this TAG image is to be associated.
The "Order" page help use re-order the channels. It can also be used to closed unused channels.
Note:
|
Closing a channel will only remove it from the mixer. It will not close the frames from the closed channel. However, since these frames are no longuer used, their associated files will show as "unselected" in the "DB File Selection (2D)" mode and can be closed there with the "Close Unselected" button. |
The "Mixer" page is where you specify how the different channels are mixed together. You can wite the equation of the pixels of the displayed image. That equation can be fairly complex and the syntax is described here. You can specify an equation for the GLI (Grey Level Image) pixels, and the TAG pixels. If the equation contain a "slider", these will be displayed under the equation window.
In our case, we want the GLI equation to be "Fat - Water", so we use "P1 - P2" for "Pixel of the channel 1 (Fat) - Pixels of the channel 2 (Water)". We also use "Tc" for the TAG pixel, meaning TAG value of the TAG channel.
This page also contain a "Bypass" button for each channel in the mix. Pressing a bypass button will temporarely change the equation to show only one of the channels. The "Shift 1, Shift 2..." can also be used to toggle the bypass mode for the different series. These shortcut keys will work even if the "mixer" interface is not displayed.
|
|
|
|
|
|
|
Channel 1 (Fat) bypass |
Channel 2 (Water) bypass |
For example, in the "Step 4: Segmentation 1 (Snake)", the script will have used the bypass for the Fat image since it is easier to segment the bone from the that images than form the mixed ones. In "Step 5: Segmentation 2 (Histogram)" the bypass are turned off and we segment with the Mixed images.