Catálogo de publicaciones - libros

Least mean square (LMS) adaptive filters have been used in a wide range of one-dimensional signal processing applications. Recently adaptive filtering are presented that are based on the Euclidean Direction Search (EDS) method of optimization. The fast version of this class is called the Fast-EDS or FEDS algorithm. The FEDS based algorithms have a fast convergence rate and O(N) computational complexity. For two-dimensional image-processing applications there is two-dimensional least mean square (TDLMS) method. This paper discusses the results of applying a TDLMS, two dimensional normalized LMS and the new two dimensional Fast Euclidean Direction Search (TDF EDS) adaptive line enhancer for the Restoration of an image contaminated by noise. The results show that the TDFEDS algorithm can follow changes in image statistics and produces a very small amount of image distortion.

The paper describes a newly presented hybrid method for high efficiency in script image feature extraction. The recognition rate was about 82% for very large number of scripts per class. However, it has reached even 100% in some cases with a smaller number of scripts per class. The system contains two projection-based methods for image characteristics extraction presented by very simple feature vectors and one image descriptor. A specially worked out thinning algorithm for the recognition system has simplified the feature extracting procedure as it provides a continuous one-pixel width skeleton of the script, which is essential for the simple-projection approach.

This paper proposes a procedure for representing image functions by a computation in two layers. It is recalled that the general function representation needs more layers than two, using the Stone-Weierstrass theorem for approximation in three layers, and the Kolmogorov theorem for representation in four layers. For achieving representation in two layers only, the requirement on a continuous representation has to removed. The Sprecher construction presented here is a general procedure for yielding such a representation in two layers. It can be used to compress images, to represent pixels and their neighborhoods directly, or to represent image operators.

The paper describes a novel behavior-based anti-Spam technology at email service based on an immune-inspired clustering algorithm. Compared with popular client anti-Spam filtering system based on content classification technology, our approach is capable of continuously delivering the most relevant Spam from the collection of all Spam that is reported by members of the network., then mail servers shall implement anti-Spam technology by using the “Black lists” that have been recognized. Experiment are discussed with real-world datasets, the conclusion have shown the technology is reliable, efficient and scalable, because no single technology can achieve one hundred percent Spam detection with zero false positives, however, it can be used in conjunction with other filtering systems to minimize errors.

In this paper, we present an efficient and biologically inspired clustering model for anomaly intrusion detection. The proposed model called Ant Colony Clustering Model (ACCM) that improves existing ant-based clustering model in searching for optimal clustering heuristically. Experimental results on KDD-Cup99 benchmark data show that ACCM is effective to detect known and unseen attacks with high detection rate and low false positive rate.

This paper describes the initial results of a research program that is designed to enhance the security of wireless mobile ad hoc networks (MANET) by developing a distributed intrusion detection capability. Our approach uses self-organizing spiking neural networks that have the ability to establish connections across a widely distributed, and highly dynamic network. This capability enables our approach to demonstrate a distributed reasoning functionality that facilitates the detection of complex attacks against MANETs. The results of the evaluation of our approach and a discussion of additional areas of research is presented.

In this paper, we verified the effect of the Distributed Visitors Coordination System in the theme park problem. The system solves the problem caused by time delay between decision-making and emergence of its effect. Its problem can happen generally in the area of congestion information system and causes the oscillation of the queue length in each service facility. In our system, some users register their next destination into a system server, the server estimates future state based on that information. Moreover, we made the exclusive queue for the agents registering next destination in each service facility to stimulate users’ registration. According to the computer simulation, even if there were many agents registering excessively, the whole waiting time reduced greatly and the oscillation of queue length in each attraction moderated.

We have been developing a new reinforcement learning called BRL, which is especially effective to multi-robot systems (MRS). BRL has a unique feature that it not only learns in the learning space but also changes the segmentation of the learning space simultaneously. BRL has been proved to be clearly effective than the other standard RL algorithms to MRS problems where the learning environment is naturally dynamic. However, we have also noticed that MRS needs the more robustness for the learning mechanism as the complexity level increases. In this paper, BRL is extended to improve the robustness against the dynamics in a learning environment by showing a way of overcoming the unwanted feature of over-fitting. Computer simulations are conducted to illustrate the robust performance of the proposed technique.

In two-sided matching problem, there are overwhelming evidences that support peoples are also motivated by concerns for fairness and reciprocity. We will show that compromise which is individually irrational improves the welfare of the whole groups. The reasonable compromise level is obtained as the function of the size of the group so that the social utility should be maximized.

Recently, tourism activity analysis has been required in order to construct effective tourism policy and strategy for personal tour which became the mainstream of Japan. In conventional tourism activity analysis, the information for analysis, i.e., tourism activity information, is collected on each tourist by using questionnaire, and then the analysis is carried out on the basis of it. However, it is difficult to realize effective tourism activity analysis in conventional method because the questionnaire-based information collection has difficulty in collecting accurate and detailed tourism activity information. In this paper, we propose a GPS log mining method in order to implement effective tourism activity analysis. GPS can automatically and continuously collect the position information where the GPS receiver exists. It is likely that an effective tourism activity analysis can be realized by extracting accurate tourism activity information from the information which is collected by GPS. We confirm the effectiveness of the proposed method through the experiments using the GPS log data collected from actual tourists in Hokkaido.