A nonlinear inverse filter is proposed for restoring signals degraded by a linear system and additive Gaussian noise. The proposed filter consists of combination of a linear high pass filter and an ε-filter, which is modified from the cascaded linear filter. The nonlinear property of the ε-filter is utilized to suppress pre-enhanced additive random noise and to restore sharp edges. It is demonstrated that the filter can be reduced to a multi-layered neural network model, and the optimal design is described by using the back propagation algorithm. The nonlinear function is approximated by a piecewise linear function, which results in simple and robust training algorithm. An application to image restoration is also presented, illustrating the effectiveness over the linear filter, especially when the amplitude of additive noise is small.

It is proposed a novel method that optimizes nonlinear filters by unsupervised learning using a novel definition of morphological pattern spectrum, called "morphological opening/closing spectrum (MOCS)." The MOCS can separate smaller portions of image objects from approximate shapes even if the shapes are degraded by noisy pixels. Our optimization method analogizes the linear low-pass filtering and Fourier spectrum: filter parameters are adjusted to reduce the portions of smaller sizes in MOCS, since they are regarded as the contributions of noises like high-frequency components. This method has an advantage that it uses only target noisy images and requires no example of ideal outputs. Experimental results of applications of this method to optimization of morphological open-closing filter for binary images are presented.

Recently, researchers in various fields have shown interest in the behavior of creatures from the viewpoint of adaptiveness and flexibility. Ants, known as social insects, exhibit collective behavior in performing tasks that can not be carried out by an individual ant. In ant colonies, chemical substances, called pheromones, are used as a way to communicate important information on global behavior. For example, ants looking for food lay the way back to their nest with a specific type of pheromone. Other ants can follow the pheromone trail and find their way to baits efficiently. In 1991, Colorni et al. proposed the ant algorithm for Traveling Salesman Problems (TSPs) by using the analogy of such foraging behavior and pheromone communication. In the ant algorithm, there is a colony consisting of many simple ant agents that continuously visit TSP cities with opinions to prefer subtours connecting near cities and they lay strong pheromones. The ants completing their tours lay pheromones of various intensities with passed subtours according to distances. Namely, subtours in TSP tourns that have the possibility of being better tend to have strong pheromones, so the ant agents specify good regions in the search space by using this positive feedback mechanism. In this paper, we propose a multiple ant colonies algorithm that has been extended from the ant algorithm. This algorithm has several ant colonies for solving a TSP, while the original has only a single ant colony. Moreover, two kinds of pheromone effects, positive and negative pheromone effects, are introduced as the colony-level interactions. As a result of colony-level interactions, the colonies can exchange good schemata for solving a problem and can maintain their own variation in the search process. The proposed algorithm shows better performance than the original algorithm with almost the same agent strategy used in both algorithms except for the introduction of colony-level interactions.

As the nonlinear adaptive filter, the neural filter is utilized to process the nonlinear signal and/or system. However, the neural filter requires large number of iterations for convergence. This letter presents a new structure of the multi-layer neural filter where the orthonormal transform is introduced into all inter-layers to accelerate the convergence speed. The proposed structure is called the transform domain neural filter (TDNF) for convenience. The weights are basically updated by the Back-Propagation (BP) algorithm but it must be modified since the error back-propagates through the orthogonal transform. Moreover, the variable step size which is normalized by the transformed signal power is introduced into the BP algorithm to realize the orthonormal transform. Through the computer simulation, it is confirmed that the introduction of the orthonormal transform is effective for speedup of convergence in the neural filter.

This paper describes a dual-looped PLL architecture to improve voltage-to-frequency linearity of VCO. The V-I converter employing a current-pumping algorithm is proposed to enhance the linearity of the VCO circuit. The designed VCO operates at a wide frequency range of 75.8 MHz-1 GHz with a good linearity. The PFD circuit design technique preventing fluctuation of the charge pump circuit under the locked condition is discussed. Simulation results show that a locking time of the proposed PLL is 3.5 µs at 1 GHz and the power dissipation is 92 mW.

In this paper, we present two parallel algorithms for computing the all nearest neighbors of an n n binary image on the Bulk-Synchronous Parallel(BSP) model. The BSP model is an asynchronous parallel computing model, where its communication features are abstracted by two parameters L and g: L denotes synchronization periodicity and g denotes a reciprocal of communication bandwidth. We propose two parallel algorithms for the all nearest neighbor problems based on two distance metrics. The first algorithm is for Lp distance, and the second algorithm is for weighted distance. Both two algorithms run in O(n2/p + L) computation time and in O(g(n/p) + L) communication time using p (1 p n) processors and in O(n2/p + (d+L)(log(p/n)/log(d+1))) computation time and in O(g(n/p) + (gd+L)(log(p/n)/log(d+1))) communication time using p (n< p n2) processors on the BSP model, for any integer d(1 dp/n).

The genetic algorithm is used to synthesize the directional circular arc array as a sectored antenna. Then, the performance of this sectored antenna in indoor wireless millimeter wave channel is investigated. Based on the desired pattern and the topography of the antennas, the synthesis problem can be reformulated into an optimization problem and solved by the genetic algorithm. The genetic algorithm will always converge to global extreme instead of local extreme and achieves a good approximation to the desired pattern. Next, the impulse responses of the indoor channel for any transmitter-receiver location are computed by shooting and bouncing ray/image techniques. By using the impulse response of multipath channel, the performance of the sectored antenna on BPSK (binary phase shift keying) system with phase and timing recovery circuits is presented. Numerical results show that the synthesized sectored antenna is effective to combat the multipath fading and can increase the transmission rate of indoor millimeter wave system.

The performance of TCP/IP over ATM over an asymmetric digital subscriber line (ADSL) was investigated. Because the bandwidth of an ADSL link can vary over time due to changes in the link's physical conditions, which degrades TCP performance, we performed simulations for various ATM traffic controls, including available bit rate (ABR) and generic flow control, used to handle variations in the ADSL bandwidth. This analysis showed that using an ABR control is effective under various traffic conditions. An ABR switch algorithm that can achieve good performance under any condition was investigated.

This paper describes the design principles, the specification, and evaluations of a new 128-bit block cipher E2, which was proposed to the AES (Advanced Encryption Standard) candidates. This algorithm supports 128-bit, 192-bit, and 256-bit secret keys. The design philosophy of E2 is highly conservative; the structure uses 12-round Feistel as its main function whose round function is constructed with 2-round SPN structure, and initial/final transformational functions. E2 has practical security against differential attack, linear attack, cryptanalysis with impossible differential, truncated differential attack, and so on. Furthermore, E2 can be implemented efficiently and flexibly on various platforms because the primitive operations involve byte length processing.

Super-anomalous elliptic curves over a ring Z/nZ ;(n=Πi=1k piei) are defined by extending anomalous elliptic curves over a prime filed Fp. They have n points over a ring Z/nZ and pi points over Fpi for all pi. We generalize Satoh-Araki-Smart algorithm and Ruck algorithm, which solve a discrete logarithm problem over anomalous elliptic curves. We prove that a "discrete logarithm problem over super-anomalous elliptic curves" can be solved in deterministic polynomial time without knowing prime factors of n.

The MOV and FR algorithms, which are representative attacks on elliptic curve cryptosystems, reduce the elliptic curve discrete logarithm problem (ECDLP) to the discrete logarithm problem in a finite field. This paper studies these algorithms and introduces the following three results. First, we show an explicit condition under which the MOV algorithm can be applied to non-supersingular elliptic curves. Next, by comparing the effectiveness of the MOV algorithm to that of the FR algorithm, it is explicitly shown that the condition needed for the MOV algorithm to be subexponential is the same as that for the FR algorithm except for elliptic curves of trace two. Finally, a new explicit reduction algorithm is proposed for the ECDLP over elliptic curves of trace two. This algorithm differs from a simple realization of the FR algorithm. Furthermore, we show, by experimental results, that the running time of the proposed algorithm is shorter than that of the original FR algorithm.

A new efficient two-dimensional warping algorithm is presented, in which sub-optimal warping is attained by iterating DP-based local optimization of warp on partially overlapping subplane sequence. From an experimental comparison with a conventional approximation algorithm based on beam search DP, relative superiority of the proposed algorithm is established.

We introduce an efficient interpolation attack which gives the tighter upper bound of the complexity and the number of pairs of plaintexts and ciphertexts required for the attack. In the previously known interpolation attack there is a problem in that the required complexity for the attack can be overestimated. We solve this problem by first, finding the actual number of coefficients in the polynomial used in the attack by using a computer algebra system, and second, by finding the polynomial with fewer coefficients by choosing the plaintexts. We apply this interpolation attack to the block cipher SNAKE and succeeded in attacking many ciphers in the SNAKE family. When we evaluate the resistance of a block cipher to interpolation attack, it is necessary to apply the interpolation attack described in this paper.

A multilayer perceptron is usually considered a passive learner that only receives given training data. However, if a multilayer perceptron actively gathers training data that resolve its uncertainty about a problem being learnt, sufficiently accurate classification is attained with fewer training data. Recently, such active learning has been receiving an increasing interest. In this paper, we propose a novel active learning strategy. The strategy attempts to produce only useful training data for multilayer perceptrons to achieve accurate classification, and avoids generating redundant training data. Furthermore, the strategy attempts to avoid generating temporarily useful training data that will become redundant in the future. As a result, the strategy can allow multilayer perceptrons to achieve accurate classification with fewer training data. To demonstrate the performance of the strategy in comparison with other active learning strategies, we also propose an empirical active learning algorithm as an implementation of the strategy, which does not require expensive computations. Experimental results show that the proposed algorithm improves the classification accuracy of a multilayer perceptron with fewer training data than that for a conventional random selection algorithm that constructs a training data set without explicit strategies. Moreover, the algorithm outperforms typical active learning algorithms in the experiments. Those results show that the algorithm can construct an appropriate training data set at lower computational cost, because training data generation is usually costly. Accordingly, the algorithm proves the effectiveness of the strategy through the experiments. We also discuss some drawbacks of the algorithm.

This paper proposes an associative memory neural network whose limiting state is the nearest point in a polyhedron from a given input. Two implementations of the proposed associative memory network are presented based on Dykstra's algorithm and a fixed point theorem for nonexpansive mappings. By these implementations, the set of all correctable errors by the network is characterized as a dual cone of the polyhedron at each pattern to be memorized, which leads to a simple amplifying technique to improve the error correction capability. It is shown by numerical examples that the proposed associative memory realizes much better error correction performance than the conventional one based on POCS at the expense of the increase of necessary number of iterations in the recalling stage.

In this paper, we present a new approach which is spanning tree-based genetic algorithm for solving a multi-objective transportation problem. The transportation problem as a special type of the network optimization problems has the special data structure in solution characterized as a transportation graph. In encoding transportation problem, we introduce one of node encodings based on a spanning tree which is adopted as it is capable of equally and uniquely representing all possible basic solutions. The crossover and mutation were designed based on this encoding. Also we designed the criterion that chromosome has always feasibility converted to a transportation tree. In the evolutionary process, the mixed strategy with (µ+λ)-selection and roulette wheel selection is used. Numerical experiments show the effectiveness and efficiency of the proposed algorithm.

A coterie is a set of quorums such that any two quorums intersect each other, and is used in a quorum based algorithm for solving the mutual exclusion problem. The availability of a coterie is the probability that the algorithm (adopting the coterie) tolerates process and/or link failures. Constructing an optimal coterie in terms of the availability is therefore important from the view of fault tolerance, but unfortunately, even calculating the availability is known to be #P-hard. Recently Harada and Yamashita proposed several heuristic methods for improving the availability of a coterie. This letter first evaluates their performance and then proposes a practical method for constructing a semi-optimal coterie by using one of the heuristic methods as a main component.

The design of array processors for solving linear systems using two-step division-free Gaussian elimination method is considered. The two-step method can be used to improve the systems based on the one-step method in terms of numerical stability as well as the requirements for high-precision. In spite of the rather complicated computations needed at each iteration of the two-step method, we develop an innovative parallel algorithm whose data dependency graph meets the requirements for regularity and locality. Then we derive two-dimensional array processors by adopting a systematic approach to investigate the set of all admissible solutions and obtain the optimal array processors under linear time-space scheduling. The array processors is optimal in terms of the number of processing elements used.

Adaptability evaluation of software systems is one of the key concerns in both software engineering and requirements engineering. In this paper, we present a formal and systematic approach to evaluate adaptability of software systems to requirements in enterprise business applications. Our approach consists of three major parts, that is, the common modeling method for both business realms and software realms, functional adaptability evaluation between the models with Σ algebra and behavioral adaptability evaluation with process algebra. By our approach, one can rigorously and uniquely determine whether a software system is adaptable to the requirements, either totally or partially. A sample application from an order processing is illustrated to show how this approach is effective in solving the adaptability evolution problem.

This paper introduces an acoustic echo canceller system materialized with a 16-bit fixed point processing type DSP (Analog Devices, ADSP-2181). This experimental system uses the tri-quantized-x individually normalized least mean square (INLMS) algorithm little degrading the convergence property under the fixed point processing. The experimental system also applies a small step gain to the algorithm to prevent the double-talk from increasing the estimation error. Such a small step gain naturally reduces the convergence speed. The experimental system compensates the reduction by applying the block length adjustment technique to the algorithm. This technique enables to ceaselessly update the coefficients of the adaptive filter even when the reference signal power is low. The experimental system thus keeps the echo return loss enhancement (ERLE) high against the double-talk.