We study methods for speeding up classification time of kernel-based classifiers. Existing solutions are based on explicitly seeking sparse classifiers during training, or by using budgeted versions of the classifier where one directly limits the number of basis vectors allowed. Here, we propose a more flexible alternative: instead of using the same basis vectors over the whole feature space, o...

In this paper, we develop numerical methods based on the weighted Birkhoff average for studying two-dimensional invariant tori volume-preserving maps. The do not rely symmetries, such as time-reversal symmetry, nor approximating by periodic orbits. rate of convergence gives a sharp distinction between chaotic and regular dynamics allows accurate computation rotation vectors Resonant rotational ...

In this paper a method for digital image orientation recognition is proposed. Feature vectors are chosen to be flip-invariant to effectively classify images onto portrait-oriented and landscapeoriented. A new texture feature is proposed based on the observation that more textured areas are located usually in the lower part of the image. The method was implemented in software and tested using an...

The present research work proposes a new approach to the model-reduction problem for nonlinear dynamical systems in the discrete-time domain based on the notion of invariant manifold. In particular, we consider nonlinear discrete-time “input-driven” real analytic dynamical systems of a structure reminiscent of skew-symmetric systems. It is implicitly assumed, that the discrete-time nonlinear dy...

The problem of finding the optimal set of quantized coefficients for a frame-based encoded signal is known to be of very high complexity. This paper presents an efficient method of finding the operational Rate-Distortion (RD) optimal set of coefficients. The major complexity reduction lies in the reformulation of the original RD-tradeoff problem, where a new set of coefficients is used as decis...

The problem of finding the optimal set of quantized coefficients for a frame-based encoded signal is known to be of very high complexity. This paper presents an efficient method of finding the operational Rate-Distortion (RD) optimal set of coefficients. The major complexity reduction lies in the reformulation of the original RD-tradeoff problem, where a new set of coefficients is used as decis...

Affine Differential Geometry provides a set of measures invariant under a larger set of transformations compared to rigid motions. This leads to several applications using robust shape descriptors. Although affine invariant operations are already used for surfaces, they do not intend to approximate the definitions of Affine Differential Geometry, which are the basis for further differential inv...

In this paper we introduce a novel domain-invariant covariance normalization (DICN) technique to relocate both in-domain and out-domain i-vectors into a third dataset-invariant space, providing an improvement for out-domain PLDA speaker verification with a very small number of unlabelled in-domain adaptation i-vectors. By capturing the dataset variance from a global mean using both development ...

The purpose of this paper is to inquire the connection between maximal inner interval estimates of the solution sets to interval linear system and solutions of the dualization equation in Kaucher interval arithmetic. The results of our work are as follows: 1) a criterion of inner interval estimate of the solution set, 2) a criterion for a solution of dualization equation to be a maximal inner i...

We establish a correspondence between the invariant subsets of a non-degenerate symmetric set-theoretical solution of the quantum Yang-Baxter equation and the parabolic subgroups of its structure group, equipped with its canonical Garside structure. Moreover, we introduce the notion of a foldable solution, which extends the one of a decomposable solution. AMS Subject Classification: 16T25, 20F36.