| genre | Resample |
| status | stable |
| author | Florian Link |
| package | MeVisLab/Standard |
| dll | MLResample1 |
| definition | MLResample1.def |
| see also | MPR, MPRPath, OrthoMPR, Reformat |
The module MPRLight calculates a multiplanar reformatted image from its input image.
The reformatting requires sampling of the input image, which can be done with different filters. The MPR is optimized (using fixpoint arithmetic) for int8 and int16 types, but has a fallback implementation for the other basic datatypes.
Depending on the Memory Access mode, it uses different strategies to access the input image.
In contrast to the module MPR, this module does not offer an Open Inventor output and is not meant to be used in an interactive context.
Provides the MPR image slice (or slice stack if Output Depth > 1).
| Aspect Ratio: Float | Memory Access: Enum | Translation: Vector3 |
| Axial: Trigger | Output Depth: Integer | |
| Coronal: Trigger | Output Size: Integer | |
| Enable Plane Edit: Bool | Plane: Plane | |
| Field Of View: Float | Represent Input Slice: Trigger | |
| Fill Value: Double | Rotation: Rotation | |
| Force Upright: Bool | Sagittal: Trigger | |
| Interpolation: Enum | Slice No.: Integer |
Sets the rotation of the MPR slice, mapping the world X-, Y-, and Z-axes to the MPR image's X-, negative Z-, and Y-axes, respectively.
The easiest way to edit the rotation is to use an OpenInventor dragger and connect its rotation field to this parameter.
Sets the translation vector to the center of the (first) MPR slice (in world coordinates).
Sets the plane containing the first slice of the MPR output image.
If checked, the Plane field can be used as an input field.
Incremental modifications of the Plane field are transformed to modifications of the Rotation and Translation parameters.
Sets the value used for regions outside of the input image.
Sets the extent of the MPR slice in the world coordinate system (in mm).
Sets the resolution of the output slice (X,Y) in pixels.
Sets the depth (Z size) of the output image.
If set to 1, a single slice is generated. If set to a value > 1, the output image is a stack of MPR's, where the translation defines the center of the first slice.
Sets the aspect ratio of the MPR.
Values larger than 1.0 produce portrait, smaller than 1.0 landscape output.
If checked, the MPR slice's Y-axis is forced to be close to the world Y-axis (or Z-axis if the MPR slice normal is parallel to the Y-axis).
Sets the slice number to be reproduced by the Represent Input Slice trigger.
If pressed, Field Of View, Output Size, Aspect Ratio, Translation, and Rotation are adjusted such that the output exactly matches the input slice number Slice No..
Defines the interpolation used for resampling.
Values:
| Title | Name | Description |
|---|---|---|
| Nearest Neighbor | NearestNeighbor | No interpolation, picks the nearest input image voxel. Fastest mode. |
| Trilinear | Trilinear | Trilinear filtering. The 8-voxel-neighborhood is used for linear interpolation. Slowest mode. |
| Trilinear Skip Border | TrilinearSkipBorder | Same as Trilinear, but all output voxels that depend on a voxel outside of the input image are filled with Fill Value. Faster than Trilinear. |
Defines how the module accesses the input image.
Values:
| Title | Name | Deprecated Name | Description |
|---|---|---|---|
| Paged | Paged | TRUE | Slower than global, except for small MPR's in large input images. The output MPR is split into tiles that are processed individually. For each tile, the module requests the corresponding 3D region of the input image. If not enough memory is available, the module falls back to Virtual Volume access. |
| Global | Global | FALSE | Fastest, but highest memory cosumption and initialization time. The input image is accessed as a global image (via the ML memory image). This consumes a lot of memory for large datasets. If not enough memory is available, the module falls back to the Virtual Volume memory access. |
| Virtual Volume | VirtualVolume | Slowest, but works on huge datasets. All pages of the input image are requested via the virtual volume mechanism, thus only small portions of the input image are needed at the same time. |