In the classical Gaussian SVM classification we use the feature space projection transforming points to normal distributions with fixed covariance matrices (identity in the standard RBF and the covariance of the whole dataset in Mahalanobis RBF). In this paper we add additional information to Gaussian SVM by considering local geometry-dependent feature space projection. We emphasize that our ap...

Although quantitative EEG parameters, such as spectral band powers, are sensitive to centrally acting drugs in dose- and time-related manners, changes of the EEG parameters are redundant. It is desirable to reduce multiple EEG parameters to a few components that can be manageable in a real space as well as be considered as parameters representing drug effects. We calculated factor loadings from...

Within the framework of kernel methods, linear data methods have almost completely been extended to their nonlinear counterparts. In this paper, we focus on nonlinear kernel techniques based on the Mahalanobis distance. Two approaches are distinguished here. The first one assumes an invertible covariance operator, while the second one uses a regularized covariance. We discuss conceptual and exp...

We propose a new image segmentation technique called strings. A string is a variational deformable model that is learned from a collection of example objects rather than built from a priori analytical or geometrical knowledge. As opposed to existing approaches, an object boundary is represented by a one-dimensional multivariate curve in functional space, a feature function, rather than by a poi...

A new biometric identification method is introduced in which a user “writes” his or her signature in the air. Accelerometers worn on a wrist device transmit acceleration data wirelessly to a host computer that processes the data and authenticates the user. This paper describes the use of curvature moments associated with 3D curves in both configuration space and velocity space as the features u...

Using the European Society of Cardiology ST-T Database, we have developed a Karhunen-Loève transform-based algorithm for robust automated detection of transient ST segment episodes during ambulatory ECG monitoring. We review current approaches and systems to detect transient ST segment changes and describe the architecture of our algorithm and its development. The algorithm incorporates a singl...

in this paper, some fundamental concepts are given relating to fuzzytopological spaces. then it is shown that there is a contravariant functorfrom the category of the pointed fuzzy topological spaces to the category ofgroups and homomorphisms. also the fuzzy topological spaces which are hopfspaces are investigated and it is shown that a pointed fuzzy toplogicalspace having the same homotopy typ...

In a fuzzy classifier with ellipsoidal regions, a fuzzy rule, which is based on the Mahalanobis distance, is defined for each class. Then the fuzzy rules are tuned so that the recognition rate of the training data is maximized. In most cases, one fuzzy rule per one class is enough to obtain high generalization ability. But in some cases, we need to partition the class data to define more than o...

Finding a good metric over the input space plays a fundamental role in machine learning. Most existing techniques assume the Mahalanobis metric without incorporating the geometry of Pn, the space of n×n symmetric positive-definite (SPD) matrices, which leads to difficulties in the optimization procedure used to learn the metric. In this paper, we introduce a novel algorithm to learn the Mahalan...

Nighttime light data from the Visible Infrared Imaging Radiometer Suite (VIIRS) Day/Night Band (DNB) provides a unique data source for mapping and monitoring urban areas at regional and global scales. This study proposes an object similarity-based thresholding method using VIIRS DNB data to map urban areas. The threshold for a target potential urban object was determined by comparing its simila...