|
|
latest version v1.9 - last update 10 Apr 2010 |
|
The parameters for the class colorACASegmentation. More...
#include <ltiColorACASegmentation.h>
Public Member Functions | |
| parameters () | |
| parameters (const parameters &other) | |
| ~parameters () | |
| const char * | getTypeName () const |
| parameters & | copy (const parameters &other) |
| parameters & | operator= (const parameters &other) |
| virtual functor::parameters * | clone () const |
| virtual bool | write (ioHandler &handler, const bool complete=true) const |
| virtual bool | read (ioHandler &handler, const bool complete=true) |
Public Attributes | |
| int | levels |
| kMColorQuantization::parameters | kMeansParam |
| float | beta |
| fvector | alpha |
| float | sigma |
| float | convergenceCriterion |
| float | tMin |
| int | nMax |
| float | windowOverlap |
| float | firstWindowSize |
| int | lastWindowSize |
| float | windowSizeStep |
The parameters for the class colorACASegmentation.
| lti::colorACASegmentation::parameters::parameters | ( | ) |
Default constructor.
Reimplemented from lti::segmentation::parameters.
| lti::colorACASegmentation::parameters::parameters | ( | const parameters & | other | ) |
Copy constructor.
| other | the parameters object to be copied |
Reimplemented from lti::segmentation::parameters.
| lti::colorACASegmentation::parameters::~parameters | ( | ) | [virtual] |
Destructor.
Reimplemented from lti::segmentation::parameters.
| virtual functor::parameters* lti::colorACASegmentation::parameters::clone | ( | ) | const [virtual] |
Returns a pointer to a clone of the parameters.
Reimplemented from lti::segmentation::parameters.
| parameters& lti::colorACASegmentation::parameters::copy | ( | const parameters & | other | ) |
Copy the contents of a parameters object.
| other | the parameters object to be copied |
Reimplemented from lti::segmentation::parameters.
| const char* lti::colorACASegmentation::parameters::getTypeName | ( | ) | const [virtual] |
Returns name of this type.
Reimplemented from lti::segmentation::parameters.
| parameters& lti::colorACASegmentation::parameters::operator= | ( | const parameters & | other | ) |
Copy the contents of a parameters object.
| other | the parameters object to be copied |
| virtual bool lti::colorACASegmentation::parameters::read | ( | ioHandler & | handler, | |
| const bool | complete = true | |||
| ) | [virtual] |
Read the parameters from the given ioHandler.
| handler | the ioHandler to be used | |
| complete | if true (the default) the enclosing begin/end will be also written, otherwise only the data block will be written. |
Reimplemented from lti::segmentation::parameters.
| virtual bool lti::colorACASegmentation::parameters::write | ( | ioHandler & | handler, | |
| const bool | complete = true | |||
| ) | const [virtual] |
Write the parameters in the given ioHandler.
| handler | the ioHandler to be used | |
| complete | if true (the default) the enclosing begin/end will be also written, otherwise only the data block will be written. |
Reimplemented from lti::segmentation::parameters.
A-Priori label membership.
If empty, all labels are equiprobable. If not empty, for those labels with an entry, this corresponds to the one-pixel clique potentials.
Default value: empty vector
Beta factor used in the Gibbs distribution model for the spatial consideration.
Default value: 0.5
Convergence criterion for a cycle.
After the mean values has been estimated, the new segmentation has to be computed using an iterative procedure that uses the previous segmentation as suggestion and continues until less than the given percentage of totall image pixels change.
Note that this value is usually much smaller than one.
Default value: 0.0004f
Initial window size as percentage of the minimum of the image dimensions.
Default value: 1.0f
Parameters for the kMeans quantization.
The number of means used in the whole algorithm will be given by the attribute kMColorQuantization::parameters::numberOfColors.
Default value: all default values except numberOfColors, which is set to 16.
Absolute minimum window size.
The window size will be reduce until it is smaller that this size.
Default value: 7
Maximum number of iterations with the same window size.
This parameter controls how much iteration will be done with the same window size of mean-estimation + segmentation-estimation steps.
Default value: 20
Standard deviation of Gaussian white distribution assumed for the region noise.
Since the value range of the data is between 0 and 255, you can consider this value as the deviation brightness expected in each of the R,G and B color channels.
Default value: 5
Threshold minimum.
The real threshold is computed multiplying this value by the current window width.
If the number of pixels in a window that belong to the current label is less than the computed threshold, the current pixel is marked as undefined and the desition to which level it will belong will be made according to spatial conditiions only (based on a MRF).
Default value: 1.0
The positioning of the window centroids will be computing using this overlap factor.
If zero is given, the windows will be place side by side. If one is given the windows will be computed for each pixel in the image, which is very expensive.
Pappas suggests the use of 0.5f.
Default value: 0.5f
Multiplicative window size step change.
This values must be possitive and strictly smaller than one.
If you set the value outsize this range, no window iteration sizes will be done, and the algorithm will behave like a k-Means clustering (but very slow).
Default value 0.5