`Classi cation of players' model is introduced and a Metropolistype Monte Carlo algorithm for manipulating the model is presented. The model is an typical example of models for the `complex' classi cation, in which the posterior means of statistics of discrete-valued parameters (labels, indicators) are prohibitively expensive to calculate with a naive method. The existence of the local maxima o...

This section discusses the procedure for extracting scale-adapted local regions, i.e., interest points equipped with characteristic scales. The Laplacian detector extracts image regions whose locations and characteristic scales are given by scale-space maxima of the Laplace operator. The Harris detector uses the same operator for scale selection, but finds the locations of interest points as th...

In this contribution we present an automatic method to extract quasisynchronous views from a sequence of monoplane rotational coronary angiograms. Based on the work of Blondel et al. we build the image of horizontal integrals, which contains information about the vertical motion of the coronary tree. The maxima of this motion correspond to end-distole states. We extract a curve of vertical moti...

This paper presents a method for simultaneous learning in multiagent environment to emerge the cooperative behaviors. Each agent has one policy and one action value function: the former is for action execution based on the the action value function updated in the previous stage, and the latter is for learning based on the episodes experienced by the 2-greedy method. This makes all agents behave...

Principal Component Analysis (PCA)-based pattern recognition is employed to predict the locations of dark matter halos within simulated galaxy distributions. Local maxima in the net tangential force produced by galaxy positions and ellipticities forms a feature space upon which PCA is performed to distinguish halo from non-halo maxima. Within a dimensionally reduced space large halo types are s...

A method for the linear discrimination of two classes is presented. It maximizes the Patrick-Fisher (PF) distance between the projected class-conditional densities. Since the PF distance is a highly nonlinear function, we propose a method, which searches for the directions corresponding to several large local maxima of the PF distance. Its novelty lies in a neural network transformation of the ...

We introduce a fibre tract segmentation algorithm based on the geometric coherence of fibre orientations as indicated by a streamline flow model. The inference of local flow approximations motivates a pairwise consistency measure between fibre ODF maxima. We use this measure in a recursive algorithm to cluster consistent ODF maxima, leading to the segmentation of white matter pathways. The meth...