MedicalHelper Class Reference
[Helpers]

[PREVIEW] [Medical] Utility class for medical applications. More...

#include <Medical/helpers/MedicalHelper.h>

Public Types
enum	Axis { AXIAL = 2, TRANSVERSE = 2, CORONAL = 1, SAGITTAL = 0 }
Public Member Functions
SbVec3f	getVoxelSize (const SoVolumeData *volume) const
SbVec3f	getPhysicalSize (const SoVolumeData *volume) const
SbVec3f	getDicomOrigin (const SoVolumeData *volume) const
Static Public Member Functions
static SbBool	orientView (MedicalHelper::Axis axis, SoCamera camera, const SoVolumeData volume=NULL, float slack=1.01)
static SbBool	dicomAdjustVolume (SoVolumeData *volume, SbBool useImagePosition=TRUE)
static SbBool	dicomAdjustVolume (SoVolumeData volume, SoMatrixTransform imgToPatient)
static SbBool	dicomAdjustDataRange (SoDataRange rangeNode, const SoVolumeData volume)
static SbBool	dicomCheckMonochrome1 (SoTransferFunction cmapNode, const SoVolumeData volume, SbBool forceReverse=FALSE)
static SbBool	dicomGetImagePosition (const SoVolumeData *volume, SbVec3f &imagePos)
static SbBool	dicomGetWindowCenterWidth (const SoVolumeData *volume, SbVec2f &winCW)
static SbVec2f	dicomGetWindowCenterWidth (const SoDataRange *dataRange)
static SbBool	dicomSetWindowCenterWidth (SoDataRange *dataRange, const SbVec2f &winCW)
static int	dicomFindFilesbySeries (const SbString &firstFile, std::vector< SbString > &files)
static int	dicomFindFilesbySeries (const SbString &firstFile, SbStringList &files)
static void	dollyZoom (float value, SoCamera *camera)
static SbBool	setFilenameList (SoVRDicomFileReader &reader, const std::vector< SbString > &list)
static SoMemoryDataAdapter *	getImageDataAdapter (const SoVolumeData *volume)
static SoVolumeData *	getVolumeData (const SoImageDataAdapter *adapter)
static SoSeparator *	buildSliceOrientationMarkers (const SoOrthoSlice *orthoSlice)
static SoSeparator *	buildSliceScaleBars (const SoCamera *camera)
static SoSeparator *	buildSliceAnnotation (const SoCamera camera, const SoOrthoSlice sliceNode, const SbString *dicomFilename)
static const SbVec2s &	exampleWindowSize ()
static SoNode *	exampleLogoNode ()
static SoNode *	exampleDicomAnnotation (const SbString &filename)
template<typename NodeType >
static NodeType *	find (SoNode *root, const SbString &nodeName=SbString(), const bool okIfNotFound=false)
template<typename NodeType >
static std::vector< NodeType * >	findNodes (SoNode *scene)
template<typename NodeType >
static NodeType *	find (SoTopLevelDialog *root, const SbString &auditorId=SbString(), bool exitIfNotFound=true)
static SbBox3f	getBoundingBox (SoNode *node)
static SoSeparator *	createCube (const SbBox3f &bbox)
static SoSeparator *	createBoundingBox (const SbBox3f &bbox, SbColor *color=NULL)
static SoSeparator *	readFile (const char *filename)
static Widget	buildInterface (Widget window, const char filename, const char viewerName, SoTopLevelDialog **topLevelDialog)

Detailed Description

[PREVIEW] [Medical] Utility class for medical applications.

Utility functions for medical visualization using Open Inventor.

Open Inventor includes a general purpose 3D graphics core library plus optimized extensions for specific data types and markets, including medical visualization. We recommend using these utility functions (and the Medical utility nodes) to get the best performance and productivity from the Open Inventor Medical Edition.

SUMMARY

dicomAdjustVolume():
Call this method to adjust VolumeViz parameters for optimal performance with medical data and to position a DICOM volume correctly in 3D space.

dicomAdjustDataRange():
Call this method to set the VolumeViz data range based on the Window Center and Window Width tags in a DICOM file. If those tags are not present, automatically sets the data range to the actual range of values in the data.

dicomSetWindowCenterWidth() and dicomGetWindowCenterWidth():
In Open Inventor the range of voxel values that map onto the color map are specified as the 'min' and 'max' values of the range (see SoDataRange). These methods provide a convenient way to set and query the data range using the medical convention of 'window center' and 'window width'.

dicomFindFilesbySeries():
Call this method to build a list of DICOM image files in a directory that belong to the same series as the specified file. This list can be passed to the DICOM volume reader using the setFilenameList() method.

getImageDataAdapter():
Call this method to get an ImageViz data set (for image processing) from a VolumeViz data set. It's faster than using SoVolumeReaderAdapter.

getVolumeData():
Call this method to get a VolumeViz data set (for visualization) from an ImageViz data set. It's faster than using SoVRImageDataReader directly.

orientView():
This method is useful for orienting the camera to view a single image (a slice in VolumeViz jargon) along one of the standard axes (Axial, Coronal or Sagittal). But remember that the volume is always a 3D object to Open Inventor and you must also set the 'axis' field on the SoOrthoSlice node. See the dicomImageViewer example.

EXAMPLE

 // Data reader: Load all images in a series
 SoVRDicomFileReader* dicomReader = new SoVRDicomFileReader();
   std::vector<SbString>& fileList;
   MedicalHelper::dicomFindFilesbySeries( firstFilename, fileList );
   MedicalHelper::setFilenameList( *dicomRender, fileList );

 // Data node
 SoVolumeData* volume = new SoVolumeData();
   volData->setReader( *dicomReader );
   MedicalHelper::dicomAdjustVolume( volData ); // Adjust position, extent, ...
   root->addChild(volData);

 // Data range (from data file by default)
 SoDataRange* dataRange = new SoDataRange();
   MedicalHelper::dicomAdjustDataRange( dataRange, volData ); // Use window center/width if possible...
   root->addChild( dataRange );

EXAMPLE

 // Start a filter pipeline from a VolumeViz data set
 SoMemoryDataAdapter* adapter = MedicalHelper::getImageDataAdapter( volume );

 // Add a filter
 SoDeblurProcessing2d* filter = new SoDeblurProcessing2d();
   filter->sharpeningFactor = 2;
   filter->inImage = adapter;

Member Enumeration Documentation

enum MedicalHelper::Axis

Medical axis names.

These names are mapped to the corresponding XYZ enum value in Open Inventor.

Enumerator:

AXIAL	Axial (Z axis).
TRANSVERSE	Transverse (synonym for Axial).
CORONAL	Coronal (Y axis).
SAGITTAL	Sagittal (X axis).

Member Function Documentation

static Widget MedicalHelper::buildInterface	(	Widget	window,
		const char *	filename,
		const char *	viewerName,
		SoTopLevelDialog **	topLevelDialog
	)			`[static]`

Convenience function to read a .iv file defining a DialogViz user interface.

static SoSeparator* MedicalHelper::buildSliceAnnotation	(	const SoCamera *	camera,
		const SoOrthoSlice *	sliceNode,
		const SbString *	dicomFilename
	)			`[static]`

Slice viewer annotations.

This is just a convenience method and helps keep the demo/example programs consistent. Applications can use SliceScaleBar (etc) directly.

Calls the buildSliceOrientationMarkers() and buildSliceScaleBars() methods and also creates some DicomInfo nodes using tag info from the optional image filename.

static SoSeparator* MedicalHelper::buildSliceOrientationMarkers ( const SoOrthoSlice * orthoSlice ) [static]

Build a scene graph to display slice orientation markers.

This is just a convenience method and helps keep the demo/example programs consistent. Applications can use SliceOrientationMarkers directly.

Creates a SliceOrientationMarkers node with a default color and background and connects its 'axis' field to the 'axis' field of the specified slice node.

static SoSeparator* MedicalHelper::buildSliceScaleBars ( const SoCamera * camera ) [static]

Build a scene graph to display dynamic scale bars for slice viewing.

This is just a convenience method and helps keep the demo/example programs consistent. Applications can use SliceScaleBar directly.

Note that a typical length, 10 cm, is assumed.

Creates horizontal and vertical SliceScaleBar nodes with a default color and line width and sets their 'trackedCamera' field to the specified camera node.

static SoSeparator* MedicalHelper::createBoundingBox	(	const SbBox3f &	bbox,
		SbColor *	color = `NULL`
	)			`[static]`

Convenience function to draw a wireframe box representing the specified bounding box.

Default color is red (1,0,0).

static SoSeparator* MedicalHelper::createCube ( const SbBox3f & bbox ) [static]

Convenience function to draw a cube representing the specified bounding box.

Color is inherited.

static SbBool MedicalHelper::dicomAdjustDataRange	(	SoDataRange *	rangeNode,
		const SoVolumeData *	volume
	)			`[static]`

Adjust data range based on values in the DICOM file, i.e.

the window center (0028,1050) and window width (0028,1051) values. Returns true if successful. Assigns default values if necessary. If there are multiple slices, values are taken from the first slice. Volume is needed to get DICOM attributes, but is not modified.

static SbBool MedicalHelper::dicomAdjustVolume	(	SoVolumeData *	volume,
		SoMatrixTransform *	imgToPatient
	)			`[static]`

Similar to dicomAdjustVolume( SoVolumeData*, SbBool ) above but returns an SoMatrixTransform that can be used to properly locate VolumeData in patient space in case of non axis aligned acquisition:.

 ...
 SoMatrixTransform* matrixTransform = new SoMatrixTransfrom();
 dicomAdjustVolume(volumeData, matrixTransfrom)
 root->addChild(matrixTransform);
 root->addChild(volumeData);
 ...

The computed matrix embeds the image position and image orientation as described here: https://dicom.innolitics.com/ciods/ct-image/image-plane/00200037 Scaling part is embedded in VolumeData extent.

If SoVolumeData is not a DICOM volume, matrixTransform will be set to identity.

Returns true if successful.

static SbBool MedicalHelper::dicomAdjustVolume	(	SoVolumeData *	volume,
		SbBool	useImagePosition = `TRUE`
	)			`[static]`

Optimize volume data node for DICOM volumes.

Returns true if successful. The volume properties are modified (but not the data values). Do not call this method before setting the volume data node's fileName field or calling setReader.

First, adjusts the LDM tile size to optimize data loading.
Note this adjustment is valid for any data format that is not already tiled, i.e. most formats except .lda and .ldm.

Second, adjusts the volume extent in 3D to be the actual physical extent of the data set measured from outside edge to outside edge of the voxels. This is a DICOM specific adjustment.

Third, if 'useImagePosition' is true, adjusts the volume extent so the center of the first voxel is positioned at the coordinates specified in the data set's "Image position" (0020,0032) attribute. By default the DICOM volume reader sets the volume extent so the center of the volume is positioned at 0,0,0 in 3D space. This is a DICOM specific adjustment.

Note that this method doesn't support non axis aligned acquisition. In such a case, prefer using the dicomAdjustVolume( SoVolumeData*, SoMatrixTransform* ) below.

static SbBool MedicalHelper::dicomCheckMonochrome1	(	SoTransferFunction *	cmapNode,
		const SoVolumeData *	volume,
		SbBool	forceReverse = `FALSE`
	)			`[static]`

Automatically adjusts the color map (transferFunction) if the specified volume is a DICOM data set (reader is SoVRDicomFileReader) and the the Photometric Interpretation (0028,0004) value is MONOCHROME1.

Specifically, if the above conditions are true, this method reverses the values in the current color map.

Typically used as a convenience call when rendering slices. For intensity based rendering of slices, set the transfer function to the predefined color map INTENSITY. This is an intensity ramp from black to white that is appropriate for Photometric Interpretation MONOCHROME2 (the most common case). When reversed it is appropriate for P.I. MONOCHROME1.

Returns true if change was needed and was successful. Volume is needed to get DICOM attributes, but is not modified. Transfer function is modified if necessary.

static int MedicalHelper::dicomFindFilesbySeries	(	const SbString &	firstFile,
		SbStringList &	files
	)			`[static]`

Get files in DICOM series.

We recommend using the alternate version of this method. Note that SbStringList is a list of pointers to SbString objects, not a list of SbString objects.

static int MedicalHelper::dicomFindFilesbySeries	(	const SbString &	firstFile,
		std::vector< SbString > &	files
	)			`[static]`

Get files in DICOM series.

Given the path to a DICOM file, returns a list containing all the files in the same directory that are part of the same DICOM series, based on the series UID (0x0020,0x000E). File name extensions are not considered.

Returns 0 if the specified file was not found or is not a valid DICOM file.

This allows the application user to select a DICOM series in a file selection dialog box by just selecting one file instead of having to select all the files explicitly.

To load the series in Open Inventor, create an SoVRDicomFileReader and call the setFilenameList() method with the file list. Then set this reader in the SoVolumeData node using the setReader() method. Note that SbStringList is a list of pointers to SbString objects, not a list of SbString objects.

The file names in the returned list are not in any particular order. It doesn't matter for loading the volume because the DICOM volume reader automatically sorts the images by their 'Slice Location' attribute (0x0020,0x1041).

See also: SbFileHelper::listFiles(). This method returns a list of files based on a specified pattern (regular expression) for the filenames.

static SbBool MedicalHelper::dicomGetImagePosition	(	const SoVolumeData *	volume,
		SbVec3f &	imagePos
	)			`[static]`

Get the "Image Position (Patient)" attribute (0020,0032) from a DICOM volume.

DICOM calls this the upper left hand corner of the image, but more precisely it's the center of the first voxel, in millimeters (mm). For VolumeViz the upper left hand corner of the image is literally the corner of the voxel, one-half voxel different from Image Position. Returns true if the query is successful. If not successful, imagePos is 0,0,0. If there are multiple slices, value is taken from the first slice.

Note the SoVRDicomData getPosition() method does not return the value of the Image Position attribute. It returns a value computed from Image Position.

static SbVec2f MedicalHelper::dicomGetWindowCenterWidth ( const SoDataRange * dataRange ) [static]

Get the window center (level) and width values from an SoDataRange node.

Returns 0,0 on failure. Convenient for displaying data range in medical terms.

static SbBool MedicalHelper::dicomGetWindowCenterWidth	(	const SoVolumeData *	volume,
		SbVec2f &	winCW
	)			`[static]`

Get the window center (level) and width values from a DICOM volume.

If the query fails, returns false and sets 'winCW' to 0,0.

Uses the Window Center (0028,1050) and Window Width (0028,1051) tags from the first image in the stack. If these tags do not exist in the volume, then window center and width are computed from the actual data min and max values. (Note that querying the actual min and max values may take some time because every voxel must be loaded.)

This method may be called with a non-DICOM volume. In that case the actual min and max data values are used.

static SbBool MedicalHelper::dicomSetWindowCenterWidth	(	SoDataRange *	dataRange,
		const SbVec2f &	winCW
	)			`[static]`

Set an SoDataRange node from the window center (level) and width values.

Returns true if successful. Convenient for converting from medical user interface to data range.

static void MedicalHelper::dollyZoom	(	float	value,
		SoCamera *	camera
	)			`[static]`

Make the scene appear larger or smaller.

Provided as a convenience to simplify application code. Effect is the same as creating an SoCameraInteractor and calling the dolly() method, but note that the value expected by that method is the inverse which is less intuitive. Using this method, values greater than 1 make the scene appear larger and values less than 1 make the scene appear smaller. For example a of 2 will make the scene approximately 2 times larger. The specific effect on the camera is a 'dolly' or a 'zoom' depending on the type of camera.

This is often useful after calling ViewAll(). That method sets the camera parameters based on a bounding sphere around the scene which often leaves the scene not "filling" the viewport.

Details:

SoPerspectiveCamera: changes the camera 'position' field. For example, values greater than 1 move the camera closer to the focal point (divide the camera's distance from the focal point by the given value), which makes the scene appear larger. This is a 'dolly'.
SoOrthographicCamera: changes the camera 'height' field. For example, values greater than 1 decrease the view volume height (scale the height by the inverse of the given value), which makes the scene appear larger. This is a 'zoom' operation.
No effect if value is <= 0 or camera is null.

static SoNode* MedicalHelper::exampleDicomAnnotation ( const SbString & filename ) [static]

Returns a collection of DicomInfo annotation nodes for Open Inventor medical examples.

This method is not important for customer appications, but the technique can be useful for adding annotation text to an application window.

filename is the path to a DICOM file from which info will be extracted.

static SoNode* MedicalHelper::exampleLogoNode ( ) [static]

Returns a logo image node for Open Inventor medical examples.

This method is not important for customer applications, but the technique can be useful for adding logo images to an application window. Returns an SoImageBackground node, but as an SoNode. The returned node has a ref count of zero.

static const SbVec2s& MedicalHelper::exampleWindowSize ( ) [static]

Returns standard window size for Open Inventor medical examples.

This method is not important for customer applications.

template<typename NodeType >

static NodeType* MedicalHelper::find	(	SoTopLevelDialog *	root,
		const SbString &	auditorId = `SbString()`,
		bool	exitIfNotFound = `true`
	)			`[inline, static]`

Convenience function to find first specified node in specified GUI graph.

template<typename NodeType >

static NodeType* MedicalHelper::find	(	SoNode *	root,
		const SbString &	nodeName = `SbString()`,
		const bool	okIfNotFound = `false`
	)			`[inline, static]`

Convenience function to find first specified node in specified scene graph.

template<typename NodeType >

static std::vector<NodeType*> MedicalHelper::findNodes ( SoNode * scene ) [inline, static]

Convenience function to find all node of specified type in specified graph.

static SbBox3f MedicalHelper::getBoundingBox ( SoNode * node ) [static]

Convenience function to retrieve bounding box of specified node.

If node is a group that contains an SoVolumeData, uses SoVolumeData extent.

SbVec3f MedicalHelper::getDicomOrigin ( const SoVolumeData * volume ) const

Returns the DICOM volume origin in 3D space.

This is just a convenience. In DICOM terms the image/volume origin is the center of the first voxel, but in VolumeViz the minimum point of the 3D extent is the outer edge of the first voxel. The difference is 1/2 voxel.

Note: dicomGetImagePosition returns the actual value of the Image Position Patient attribute from the DICOM file (if possible).

static SoMemoryDataAdapter* MedicalHelper::getImageDataAdapter ( const SoVolumeData * volume ) [static]

Make a VolumeViz data set available in ImageViz.

Creates and returns an SoMemoryDataAdapter with the same dimensions, extent and data as the SoVolumeData. Returns NULL on failure, e.g. if volume is empty.

Note that this operation makes a copy of the data set. So it must be possible to allocate a contiguous block of memory large enough to hold the whole volume. It is not currently possible for ImageViz and VolumeViz to share a data set.

You can use SoVolumeReaderAdapter to get the same result but the performance is much slower because SoVolumeReaderAdapter loads data using the volume's associated volume reader class and re-reads the data from disk. This convenience method is much faster because it loads data using the LDMDataAccess interface which takes advantage of data that VolumeViz already has in memory.

Performance depends on the 'tile size' set for the source volume. It will be slow (but still faster than SoVolumeReaderAdapter) if you use the default tile size (64) with a 512^3 volume. We recommend setting the volume's tile size equal to the volume size (or calling the dicomAdjustVolume() method which does this automatically).

SbVec3f MedicalHelper::getPhysicalSize ( const SoVolumeData * volume ) const

Returns the volume's physical size in 3D space.

This is just a convenience. The physical size in 3D is simply the span of the volume extent.

static SoVolumeData* MedicalHelper::getVolumeData ( const SoImageDataAdapter * adapter ) [static]

Make an ImageViz data set available in VolumeViz.

Creates an SoVRImageDataReader to access the data and automatically sets the SoVolumeData's 'tile size' to an optimal value. Note that this operation makes a copy of the data set. It is not currently possible for ImageViz and VolumeViz to share a data set. Returns NULL on failure.

You can use SoVRImageDataReader directly, but the resulting volume may be slow to access if you do not set the tile size and the volume dimensions are larger than the default tile dimension (64). Setting the tile size allows VolumeViz to access the data as a single tile when the dimensions are <= 512 or at least to access the data using a larger tile size (which means a smaller number of tiles).

SbVec3f MedicalHelper::getVoxelSize ( const SoVolumeData * volume ) const

Returns the voxel size in 3D space.

This is just a convenience. The voxel size is simply the volume extent in 3D space divided by the number of voxels.

static SbBool MedicalHelper::orientView	(	MedicalHelper::Axis	axis,
		SoCamera *	camera,
		const SoVolumeData *	volume = `NULL`,
		float	slack = `1.01`
	)			`[static]`

Orient view.

Adjusts the camera to view a slice or volume along the specified axis. The camera is rotated to give the conventional orientation for the specified axis and viewAll() is called to center the position and adjust the view volume.

If the camera is an SoOrthographicCamera (the usual case for viewing a slice) and the volume is specified, then the view volume height is adjusted to make the slice image fill the window. This is convenient because the standard Open Inventor viewAll() methods give a very loose fit around the slice image.

For a 3D volume rendering view, typically specify the Coronal axis.

axis: Axial, Coronal or Sagittal
camera: The camera node that will be modified (fails if null).
volume: The data volume that is being viewed (used to compute view volume, not modified). May be null, but a default volume extent of -1 to 1 will be used.
slack: Values greater than 1 will leave additional space between image and window edge.

Returns true if the operation succeeded.

static SoSeparator* MedicalHelper::readFile ( const char * filename ) [static]

Convenience function to read a .iv file and return pointer to the root of the created scenegraph.

static SbBool MedicalHelper::setFilenameList	(	SoVRDicomFileReader &	reader,
		const std::vector< SbString > &	list
	)			`[static]`

Set filename list using a list of strings (file names).

This is a convenient alternative to calling SoVRDicomFileReader's method setFilenameList() directly. That method requires an SbStringList object which is a simple list of _pointers_, not a list of objects.

The documentation for this class was generated from the following file:

Medical/helpers/MedicalHelper.h

MedicalHelper Class Reference [Helpers]

Public Types

Public Member Functions

Static Public Member Functions

Detailed Description

EXAMPLE

EXAMPLE

SEE ALSO

Member Enumeration Documentation

Member Function Documentation

MedicalHelper Class Reference
[Helpers]