S

S Verzakov

17 records found

Kernel combination versus classifier combination

Estimating grassland biomass using SVM band shaving of hyperspectral data

Augmented embedding of dissimilarity data into (pseudo-)Euclidean spaces
Pairwiseproximitiesdescribethepropertiesofobjectsintermsoftheirsimilarities.Byusingdi¿erentdistance-basedfunctionsonemayencodedi¿erentcharacteristicsofagivenproblem.However,tousetheframeworkofstatisticalpatternrecognitionsomevectorrepresentationshouldbeconstructed.Oneofthesimples ...

Augmented embedding of dissimilarity data into (pseudo-) Euclidean spaces

Edge detection in hyperspectral imaging: multivariate statistical approaches
Edgedetectioniswelldevelopedareaofimageanalysis.Manyvariouskindsoftechniquesweredesignedforone-channelimages.Also,aconsiderableattentionwaspaidtoedgedetectionincolor,multispectral,andhyperspectralimages.However,therearestillmanyopenissuesinedgedetectioninmultichannelimages.Forexa ...

Effectiveness of spectral band selection/extraction techniques for spectral data
In the past few years a variety of successful algorithms to select/extract discriminative spectral bands was introduced. By exploiting the connectivity of neighbouring spectral bins, these techniques may be more beneficial than the standard feature selection/extraction methods ap ...

Domain based LDA and QDA

Statistical pattern recognition approach to edge detection in multichannel images

Simplifying the model-based classifiers for multi-modal problems in classification of spectra

Improving the maximum-likelihood co-occurrence classifier: a study on classification of inhomogeneous rock images

Estimating spatial patterns of biomass and nitrogen status in grasslands through imaging spectrometry

The tangent kernel approach to illumination-robust texture classification

The tangent kernel SVM for calibration-stable histogram discrimination
Conference paper (2005) - S Verzakov (author), Bob Duin (author)

Multi-class extensions of the GLDB feature extraction algorithm for spectral data

Feature shaving for spectroscopic data

Exploiting connectivity in spectral data: SVM band-shaving

Cost-based classifier evaluation for imbalanced problems