Karsten Schulz

Learn More
In many practical classification problems the severeness of misclassifications depends on the semantics of true and predicted class in the underlying domain. We present such a problem from machine vision, where additionally the class probabilities are extremely unbalanced. Due to their in-terpretability neuro-fuzzy classifiers axe a popular way to extract(More)
Nearest neighbor (NN) techniques are commonly used in remote sensing, pattern recognition and statistics to classify objects into a predefined number of categories based on a given set of predictors. These techniques are especially useful in those cases exhibiting highly nonlinear relationship between variables. In most studies the distance measure is(More)
A new class of problems has been identified for workflow management wherein this technology imposes an additional overhead on users which is not adequately compensated by the benefits gained. Multiple observations from real cases identify a contradiction between the preferred work practice and some fundamental principles behind workflow systems. The focus(More)
The improved ground resolution of state-of-the-art synthetic aperture radar (SAR) sensors suggests utilizing SAR data for the analysis of urban areas. Building recognition however suffers from the consequences of the inherent side-looking viewing geometry of SAR, particularly occlusion and layover hinder the analysis. Usually extracted layover regions are(More)