Validate SVM Images Classifier Application

Brief Description

Estimate the performance of the SVM model with a new set of validation samples and another image.

Tags

Learning

Long Description

Estimate the performance of the SVM model obtained by the ImagesSVMClassifier with a new set of validation samples and another image. The application asks for images statisctics as input (XML file generated with the ComputeImagesStatistics application) and a SVM model (text file) generated with the ImagesSVMClassifier application. It will compute the global confusion matrix and kappa index and also the precision, recall and F-score of each class.

Parameters

• [param] Input Image List (-il): Input image list filename.

• [param] Vector Data List (-vd): Vector Data of samples used to validate the estimator.

• [param] DEM repository (-dem): Path to DEM repository.

• [param] XML image statistics file (-imstat): Filename of an XML file containing mean and standard deviation of input images.

• [param] Output filename (-out): Filename, which contains the performances of the SVM model.

• [param] SVM validation filename (-svm): SVM model to validate (given by TrainSVMImagesClassification output for instance).

Limitations

None

Authors

OTB-Team

See also

Example of use

• Parameters to set value:

• Input Image List: Classification/QB_1_ortho.tif

• Vector Data List: VectorData_QB1_bis.shp

• XML image statistics file: clImageStatisticsQB123.xml

• Output filename: PerformanceEstimationQB123.txt

• SVM validation filename: clsvmModelQB123.svm

• Command line to execute:

otbcli_ValidateSVMImagesClassifier -il Classification/QB_1_ortho.tif -vd VectorData_QB1_bis.shp -imstat clImageStatisticsQB123.xml -out PerformanceEstimationQB123.txt -svm clsvmModelQB123.svm