ImageDev

TextureClassificationTrain

Enriches a texture model by training it on a gray level image that has been labeled.

Access to parameter description

As an introduction: This algorithm can work on 3D, sequence, and spectral images.
Following the computation mode of the associated texture classification model (2D or 3D) initialized with TextureClassificationCreate, the input image can be considered as a series of 2D images or as a series of 3D images.

For all the labeled pixels of the training image, this algorithm extracts the texture features defined by the groups selected in the texture classification model. These newly extracted features increase the training-set and enrich the model.

Note: All the pixels within the analysis radius of a labeled pixel are considered for the feature extraction. It means that the number of pixels used for training is greater than the number of the labeled pixels.

See also

Function Syntax

This function returns inputOutputModel. 
// Function prototype
TextureClassificationModel::Ptr
textureClassificationTrain( std::shared_ptr< iolink::ImageView > inputImage,
                            std::shared_ptr< iolink::ImageView > inputTrainingImage,
                            double minSeparationPercentage,
                            bool finalizeTraining,
                            TextureClassificationModel::Ptr inputOutputModel = NULL );

Class Syntax

Parameters

Parameter Name Description Type Supported Values Default Value
input
inputImage
The input grayscale image. Image Grayscale nullptr
input
inputTrainingImage
The input label image.
Each label represents a class sample for the training. 		Image Label nullptr
input
minSeparationPercentage
This parameter controls the rejection criteria of the feature selection algorithm (FS).

A measure is rejected if its contribution does not increase the separation power of the classification model enough. This ratio indicates the minimal relative growth required to keep a measure. More information is available in the Feature Selection section.
This value must be greater than or equal to 0.0. 		Float64 [0, 100] 5
input
finalizeTraining
This parameter indicates whether the algorithm must finalize the training. This parameter must be set to true since the retraining is currently not allowed. Bool true
input and output
inputOutputModel
The input and output texture classification model. TextureClassificationModel nullptr

Object Examples

auto classification_input = readVipImage( std::string( IMAGEDEVDATA_IMAGES_FOLDER ) + "classification_input.vip" );
auto classification_input_label = readVipImage( std::string( IMAGEDEVDATA_IMAGES_FOLDER ) + "classification_input_label.vip" );
TextureClassificationModel::Ptr modelToTrain= TextureClassificationModel::read( std::string( IMAGEDEVDATA_OBJECTS_FOLDER ) + "modelToTrain.vip" );

TextureClassificationTrain textureClassificationTrainAlgo;
textureClassificationTrainAlgo.setInputImage( classification_input );
textureClassificationTrainAlgo.setInputTrainingImage( classification_input_label );
textureClassificationTrainAlgo.setMinSeparationPercentage( 5 );
textureClassificationTrainAlgo.setFinalizeTraining( true );
textureClassificationTrainAlgo.setInputOutputModel( modelToTrain );
textureClassificationTrainAlgo.execute();

std::cout << "computeMode: " << textureClassificationTrainAlgo.inputOutputModel()->computeMode( ) ;

Function Examples

auto classification_input = readVipImage( std::string( IMAGEDEVDATA_IMAGES_FOLDER ) + "classification_input.vip" );
auto classification_input_label = readVipImage( std::string( IMAGEDEVDATA_IMAGES_FOLDER ) + "classification_input_label.vip" );
TextureClassificationModel::Ptr modelToTrain= TextureClassificationModel::read( std::string( IMAGEDEVDATA_OBJECTS_FOLDER ) + "modelToTrain.vip" );

auto result = textureClassificationTrain( classification_input, classification_input_label, 5, true, modelToTrain );

std::cout << "computeMode: " << result->computeMode( ) ;