#include <itkKLMRegionGrowImageFilter.h>
Inheritance diagram for itk::KLMRegionGrowImageFilter< TInputImage, TOutputImage >:
Public Types | |
| typedef KLMRegionGrowImageFilter | Self |
| typedef RegionGrowImageFilter< TInputImage, TOutputImage > | Superclass |
| typedef SmartPointer< Self > | Pointer |
| typedef SmartPointer< const Self > | ConstPointer |
| typedef TInputImage | InputImageType |
| typedef TInputImage::Pointer | InputImagePointer |
| typedef TInputImage::ConstPointer | InputImageConstPointer |
| typedef TInputImage::PixelType | InputImagePixelType |
| typedef TInputImage::PixelType::VectorType | InputImageVectorType |
| typedef TInputImage::IndexType | InputImageIndexType |
| typedef TInputImage::OffsetType | InputImageOffsetType |
| typedef ImageRegionIterator< TInputImage > | InputImageIterator |
| typedef ImageRegionConstIterator< TInputImage > | InputImageConstIterator |
| typedef TOutputImage | OutputImageType |
| typedef TOutputImage::Pointer | OutputImagePointer |
| typedef TOutputImage::PixelType | OutputImagePixelType |
| typedef TOutputImage::IndexType | OutputImageIndexType |
| typedef TOutputImage::OffsetType | OutputImageOffsetType |
| typedef ImageRegionIterator< TOutputImage > | OutputImageIterator |
| typedef TOutputImage::PixelType::VectorType | OutputImageVectorType |
| typedef Image< unsigned short, itkGetStaticConstMacro(LabelImageDimension) | LabelImageType ) |
| typedef LabelImageType::Pointer | LabelImagePointer |
| typedef LabelImageType::PixelType | LabelImagePixelType |
| typedef LabelImageType::IndexType | LabelImageIndexType |
| typedef LabelImageType::OffsetType | LabelImageOffsetType |
| typedef ImageRegionIterator< LabelImageType > | LabelImageIterator |
| typedef vnl_matrix< double > | VecDblType |
| typedef KLMSegmentationBorder | BorderType |
| typedef KLMDynamicBorderArray< BorderType > | SegmentationBorderPtr |
Public Member Functions | |
| virtual const char * | GetClassName () const |
| itkStaticConstMacro (LabelImageDimension, unsigned int, TInputImage::ImageDimension) | |
| virtual void | SetMaxLambda (unsigned int _arg) |
| virtual unsigned int | GetMaxLambda () |
| LabelImagePointer | GetLabelledImage (void) |
| void | PrintAlgorithmRegionStats (void) |
| void | PrintAlgorithmBorderStats (void) |
| void | PrintAlgorithmBorderStats (bool smartBorderPointerUseFlag) |
| virtual void | CalculateInitRegionStats (int regionRowIndex, int regionColIndex, int regionRowGridSize, int regionColGridSize) |
| virtual void | CalculateInitRegionStats (int regionRowIndex, int regionColIndex, int regionSliceIndex, int regionRowGridSize, int regionColGridSize, int regionSliceGridSize) |
Static Public Member Functions | |
| Pointer | New () |
Protected Member Functions | |
| KLMRegionGrowImageFilter () | |
| ~KLMRegionGrowImageFilter () | |
| void | PrintSelf (std::ostream &os, Indent indent) const |
| virtual void | GenerateData () |
| virtual void | GenerateInputRequestedRegion () |
| virtual void | EnlargeOutputRequestedRegion (DataObject *) |
| virtual void | GenerateOutputInformation () |
| void | ApplyRegionGrowImageFilter () |
| virtual void | MergeRegions () |
| void | GenerateOutputImage (unsigned int imgWidth, unsigned int imgHeight) |
| void | GenerateOutputImage (unsigned int imgWidth, unsigned int imgHeight, unsigned int imgDepth) |
| void | ApplyKLM () |
| void | InitializeKLM (unsigned int imgWidth, unsigned int imgHeight) |
| void | InitializeKLM (unsigned int imgWidth, unsigned int imgHeight, unsigned int imgDepth) |
| LabelImagePointer | localfn_generate_labeled2Dimage (LabelImageType *labelImagePtr) |
| LabelImagePointer | localfn_generate_labeled3Dimage (LabelImageType *labelImagePtr) |
itkKLMRegionGrowImageFilter is the base class for the KLMRegionGrowImageFilter objects. This object performes energy-based region growing for multiband images. Since this is based on G. Koepfler,C. Lopez and J. M. Morel's work described below, the acronym KLM is added at the end of the object name The ApplyRegionGrowImageFilter() function implements the segmentation algorithm that partitions the input image into non-overlapping regions by minimizing an energy functional which trades off the similarity of regions against the length of their shared boundary. The heart of the prcess relies on the MergeRegion() method that calls a private function to perform the merging of region based on the piecewise constance KLM algorithm for region merging. For extensibiltiy purposes, the MergeRegion() function is made virtual. Extensions can be made possible using function overloading or overriding the virutual function in a derived class. It starts by breaking the image into many small regions and fitting the regions to a polynomial model. The algorithm iteratively merges into one region the two adjoining regions which are most alike in terms of the specified polynomial model given the length of the border between the two regions. Internally, the energy functional is evaluated using a Lagrangian parameter called lambda which is also called the scale parameter as it controls the coarseness of the segmentation where a small value of lambda corresponds to a finer segmentation with more regions and a large value corresponds to a coarse segmentation with fewer regions. Since the algorithm grows regions by merging like regions, the internal value of lambda increases as the number of regions decreases.
The user can stop the merging of regions using the SetMaxNumRegions() and SetLambda() functions. The SetMaxNumRegions() fuction is publicly inherited from its base class and internally sets the m_MaxNumRegions parameter. The SetLambda() function sets the m_Lambda parameter. If the number of regions in the image is equal to m_MaxNumRegions or if the internal energy functional becomes greater than m_Lambda, then the merging iterations will stop. Note that a larger value for m_Lambda will result in fewer boundaries and fewer regions, while a smaller value for m_Lambda will result in more boundaries and more regions. To have m_MaxNumRegions control exactly the number of output regions, m_Lamda should be set to a very large number. To have m_Lambda control exactly the number of output regions, m_MaxNumRegions should be set to 2. As a default value the maximum lambda value is set to 1000 and m_MaxNumRegions is set to 2 as default.
Currently implementation puts equal weight to the multichannel values. In future improvements we plan to allow the user to control the weights associated with each individual channels.
It is templated over the type of input and output image. This object supports data handling of multiband images. The object accepts images in vector format, where each pixel is a vector and each element of the vector corresponds to an entry from 1 particular band of a multiband dataset. We expect the user to provide the input to the routine in vector format. A single band image is treated as a vector image with a single element for every vector.
This algorithm implementation takes a multiband image stored in vector format as input and produces two outputs. Using the ImageToImageFilter, the piecewise constant approximation image is the output calculated using the process update mechanism. The second output, i.e., the image with the region labels (segmentation image) is returned at users request by calling GetLabelledImage() function. This function returns a reference to the labelled image determined using the KLM algorithm. The algorithm supports 2D and 3D data sets only. The input image dimensions must be exact multiples of the user specified gridsizes. Appropriate padding must be performed by the user if any image which are not mutliples of the gridsizes are used.
For more information about the algorithm, see G. Koepfler, C. Lopez and J. M. Morel, ``A Multiscale Algorithm for Image Segmentation by Variational Method,'' {SIAM Journal of Numerical Analysis}, vol. 31, pp. 282-299, 1994.
Algorithm details:
This function segments a two-dimensional input image into non-overlapping regions $O_i$, i=1,2,...,N, where N is the total number of region, by minimizing the following energy functional (also known as the simplified Mumford and Shah functional): $E(u,K)={-K}||u(r,c)-g(r,c)||^2{d{}}+{L(K)}$, where $$ denotes the domain of an image, g(r,c) is the input image, and u(r,c) is an approximation of g(r,c). Furthermore, u(r,c) is defined to be piecewise constant in regions $O_i$. If $ O_i$ represents the boundary of the region, $K={i=1}^N{O_i}$ denotes the set of all region boundaries and L(K) is the total length of the boundaries. The parameter $$ controls the coarseness of the segmentation (i.e. a larger $$ will result in fewer boundaries).
Starting with small, piecewise-constant initial regions the algorithm iteratively merges the two adjacent regions $O_i$ and $O_j$ which most decrease the energy functional. In other words, the merging criterion is based on the difference between the current energy E(u,K) and the energy that would result after a merge, $E({u},K-(O_i,O_j))$, where ${u}$ is the piecewise constant approximation of the input image g, and $(O_i,O_j)$ is the common boundary between region $O_i$ and $O_j$. It can be shown that $E(u,K)-E({u},K-(O_i,O_j))= {L((O_i,O_j))}- {(|O_i| |O_j|) (|O_i|+|O_j|)} ||c_i-c_j||^2$.
Once two regions are merged the following update equations are used to calculated the constant approximation of the new region:
$c_{i,j} = (c_i |O_i| + c_j |O_j|) (|O_i| + |O_j|)$.
Again, the merging of regions continues until the desired number of regions has been reached or until the desired coarseness (specified by the scale parameter $$) has been reached.
The two outputs are possible to derive from the object: (1) u, the piecewise constant approximation (mean of the regions) to the input image set; This is currently generated by the process object pipeline and the (2) the labelled regions in the input image set is generated by the GetLabelledImage() function.
Definition at line 157 of file itkKLMRegionGrowImageFilter.h.
|
|||||
|
Type definition for the smart border type. Definition at line 238 of file itkKLMRegionGrowImageFilter.h. |
|
|||||
|
Reimplemented from itk::RegionGrowImageFilter< TInputImage, TOutputImage >. Definition at line 164 of file itkKLMRegionGrowImageFilter.h. |
|
|||||
|
Definition at line 191 of file itkKLMRegionGrowImageFilter.h. |
|
|||||
|
Reimplemented from itk::RegionGrowImageFilter< TInputImage, TOutputImage >. Definition at line 175 of file itkKLMRegionGrowImageFilter.h. |
|
|||||
|
Type definition for the input image index type. Definition at line 184 of file itkKLMRegionGrowImageFilter.h. |
|
|||||
|
Type definition for the image iterators to be used. Definition at line 190 of file itkKLMRegionGrowImageFilter.h. |
|
|||||
|
Type definition for the input image offset type. Definition at line 187 of file itkKLMRegionGrowImageFilter.h. |
|
|||||
|
Type definition for the input image pixel type. Reimplemented from itk::RegionGrowImageFilter< TInputImage, TOutputImage >. Definition at line 178 of file itkKLMRegionGrowImageFilter.h. |
|
|||||
|
Reimplemented from itk::RegionGrowImageFilter< TInputImage, TOutputImage >. Definition at line 174 of file itkKLMRegionGrowImageFilter.h. |
|
|||||
|
Type definition for the input image. Reimplemented from itk::RegionGrowImageFilter< TInputImage, TOutputImage >. Definition at line 173 of file itkKLMRegionGrowImageFilter.h. |
|
|||||
|
Type definition for the input image pixel vector type. Definition at line 181 of file itkKLMRegionGrowImageFilter.h. |
|
|||||
|
Type definition for the output image index type. Definition at line 226 of file itkKLMRegionGrowImageFilter.h. |
|
|||||
|
Type definition for the labelled image iterators. Definition at line 232 of file itkKLMRegionGrowImageFilter.h. |
|
|||||
|
Type definition for the output image offset type. Definition at line 229 of file itkKLMRegionGrowImageFilter.h. |
|
|||||
|
Type definition for the labelled image pixel type. Definition at line 223 of file itkKLMRegionGrowImageFilter.h. |
|
|||||
|
Type definition for the labelled image pointer. Definition at line 220 of file itkKLMRegionGrowImageFilter.h. |
|
|||||
|
Type definition for the labelled image pixel type. Definition at line 217 of file itkKLMRegionGrowImageFilter.h. |
|
|||||
|
Type definition for the output image index type. Definition at line 201 of file itkKLMRegionGrowImageFilter.h. |
|
|||||
|
Type definition for the output image iterators. Definition at line 207 of file itkKLMRegionGrowImageFilter.h. |
|
|||||
|
Type definition for the output image offset type. Definition at line 204 of file itkKLMRegionGrowImageFilter.h. |
|
|||||
|
Type definition for the output image pixel type. Reimplemented from itk::RegionGrowImageFilter< TInputImage, TOutputImage >. Definition at line 198 of file itkKLMRegionGrowImageFilter.h. |
|
|||||
|
Reimplemented from itk::RegionGrowImageFilter< TInputImage, TOutputImage >. Definition at line 195 of file itkKLMRegionGrowImageFilter.h. |
|
|||||
|
Type definition for the output image. Reimplemented from itk::RegionGrowImageFilter< TInputImage, TOutputImage >. Definition at line 194 of file itkKLMRegionGrowImageFilter.h. |
|
|||||
|
Type definition for the output image pixel vector type. Definition at line 210 of file itkKLMRegionGrowImageFilter.h. |
|
|||||
|
Reimplemented from itk::RegionGrowImageFilter< TInputImage, TOutputImage >. Definition at line 163 of file itkKLMRegionGrowImageFilter.h. |
|
|||||
|
Type definition for the smart border pointers object. Definition at line 241 of file itkKLMRegionGrowImageFilter.h. |
|
|||||
|
Standard class typedefs. Reimplemented from itk::RegionGrowImageFilter< TInputImage, TOutputImage >. Definition at line 161 of file itkKLMRegionGrowImageFilter.h. |
|
|||||
|
Reimplemented from itk::RegionGrowImageFilter< TInputImage, TOutputImage >. Definition at line 162 of file itkKLMRegionGrowImageFilter.h. |
|
|||||
|
Storage type for the mean region intensity. Definition at line 235 of file itkKLMRegionGrowImageFilter.h. |
|
|||||||||
|
|
|
|||||||||
|
|
|
|||||||||
|
Function that calls the KLM region growing algorithm. |
|
|||||||||
|
This is the interface function that calls the specific algorithm implementation of region growing. Reimplemented from itk::RegionGrowImageFilter< TInputImage, TOutputImage >.
|
|
||||||||||||||||||||||||||||||||
|
Calculate the statistics representing the 3D regions. In this case we comput the mean region intensity and the volume of the initial rectangular area. This is the function that can be overriden in order to enable a different statistical representation for region initialization. |
|
||||||||||||||||||||||||
|
Calculate the statistics representing the 2D regions. In this case we compute the mean region intensity and the area of the initial rectangular area. This is the function that can be overriden in order to enable a different statistical representation for region initialization. |
|
||||||||||
|
Give the process object a chance to indictate that it will produce more output than it was requested to produce. For example, many imaging filters must compute the entire output at once or can only produce output in complete slices. Such filters cannot handle smaller requested regions. These filters must provide an implementation of this method, setting the output requested region to the size they will produce. By default, a process object does not modify the size of the output requested region. Reimplemented from itk::ProcessObject.
|
|
|||||||||
|
A version of GenerateData() specific for image processing filters. This implementation will split the processing across multiple threads. The buffer is allocated by this method. Then the BeforeThreadedGenerateData() method is called (if provided). Then, a series of threads are spawned each calling ThreadedGenerateData(). After all the threads have completed processing, the AfterThreadedGenerateData() method is called (if provided). If an image processing filter cannot be threaded, the filter should provide an implementation of GenerateData(). That implementation is responsible for allocating the output buffer. If a filter an be threaded, it should NOT provide a GenerateData() method but should provide a ThreadedGenerateData() instead.
Reimplemented from itk::ImageSource< TOutputImage >.
|
|
|||||||||
|
What is the input requested region that is required to produce the output requested region? The base assumption for image processing filters is that the input requested region can be set to match the output requested region. If a filter requires more input (for instance a filter that uses neighborhoods needs more input than output to avoid introducing artificial boundary conditions) or less input (for instance a magnify filter) will have to override this method. In doing so, it should call its superclass' implementation as its first step. Note that imaging filters operate differently than the classes to this point in the class hierachy. Up till now, the base assumption has been that the largest possible region will be requested of the input. This implementation of GenerateInputRequestedRegion() only processes the inputs that are a subclass of the ImageBase<InputImageDimension>. If an input is another type of DataObject (including an Image of a different dimension), they are skipped by this method. The subclasses of ImageToImageFilter are responsible for providing an implementation of GenerateInputRequestedRegion() when there are multiple inputs of different types.
Reimplemented from itk::ImageToImageFilter< TInputImage, TOutputImage >.
|
|
||||||||||||||||||||
|
Generate output approximated image. |
|
||||||||||||||||
|
Generate output approximated image. |
|
|||||||||
|
Generate the information decribing the output data. The default implementation of this method will copy information from the input to the output. A filter may override this method if its output will have different information than its input. For instance, a filter that shrinks an image will need to provide an implementation for this method that changes the spacing of the pixels. Such filters should call their superclass' implementation of this method prior to changing the information values they need (i.e. GenerateOutputInformation() should call Superclass::GenerateOutputInformation() prior to changing the information. Reimplemented from itk::ProcessObject.
|
|
|||||||||
|
Run-time type information (and related methods). Reimplemented from itk::RegionGrowImageFilter< TInputImage, TOutputImage >.
|
|
||||||||||
|
Generate labelled image. |
|
|||||||||
|
Get the maximum lambda value set by the user. |
|
||||||||||||||||||||
|
Initialize the RegionGrowImageFilter algorithm (2D case) |
|
||||||||||||||||
|
Initialize the RegionGrowImageFilter algorithm (2D case). |
|
||||||||||||||||||||
|
The dimension of the labeled image. |
|
||||||||||
|
Generate the labeled image for a 2D image. |
|
||||||||||
|
Generate the labeled image for a 3D image. |
|
|||||||||
|
Function responsible for merging two regions using energy-based regions growing criteria until the desired number of regions has been reached. When merging two regions, the smaller label is always assigned to the new region. This is consistent with the connected components algorithm. Reimplemented from itk::RegionGrowImageFilter< TInputImage, TOutputImage >.
|
|
|||||||||
|
Method for creation through the object factory. Reimplemented from itk::RegionGrowImageFilter< TInputImage, TOutputImage >.
|
|
||||||||||
|
Function that prints all the border information. |
|
||||||||||
|
Function that prints all the border information. |
|
||||||||||
|
Function that prints all the region information. |
|
||||||||||||||||
|
Methods invoked by Print() to print information about the object including superclasses. Typically not called by the user (use Print() instead) but used in the hierarchical print process to combine the output of several classes. Reimplemented from itk::RegionGrowImageFilter< TInputImage, TOutputImage >.
|
|
||||||||||
|
Set the desired threshold parameter for lambda. See itkSegmentationBorder documentation for details regarding this parameter. |
1.3.8 written by Dimitri van Heesch,
© 1997-2000