This paper proposes a design method of feature spaces in a two-stage image recognition method that improves the recognition accuracy and efficiency in statistical image recognition. The two stages are (1) image screening and (2) image recognition. Statistical image recognition methods require a lot of calculations for spatially matching between subimages and reference patterns of the specified objects to be detected in input images. Our image screening method is effective in lowering the calculation load and improving recognition accuracy. This method selects a candidate set of subimages similar to those in the object class by using a lower dimensional feature vector, while rejecting the rest. Since a set of selected subimages is recognized by using a higher dimensional feature vector, overall recognition efficiency is improved. The classifier for recognition is designed from the selected subimages and also improves recognition accuracy, since the selected subimages are less contaminated than the originals. Even when conventional recognition methods based on linear transformation algorithms, i. e. principal component analysis (PCA) and projection pursuit (PP), are applied to the recognition stage in our method, recognition accuracy and efficiency may be improved. A new criterion, called a screening criterion, for measuring overall efficiency and accuracy of image recognition is introduced to efficiently design the feature spaces of image screening and recognition. The feature space for image screening are empirically designed subject to taking the lower number of dimensions for the feature space referred to as LS and the larger value of the screening criterion. Then, the recognition feature space which number of dimensions is referred to as LR is designed under the condition LSLR. The two detection tasks were conducted in order to examine the performance of image screening. One task is to detect the eye- and-mouth-areas in a face image and the other is to detect the text-area in a document image. The experimental results demonstrate that image screening for these two tasks improves both recognition accuracy and throughput when compared to the conventional one-stage recognition method.

Mobile databases will play an important role in mobile computing environment, to provide data storing and data retrieval functionalities which are needed by most applications. In mobile computing environment, the wireless communication poses some problems, which require us to minimize its use. Replication is a database technique that is commonly used to fulfill the requirement in minimizing network usage. In this paper, we propose two replica allocation strategies, called primary-copy tracking replica allocation (PTRA) and user majority replica allocation (UMRA), which are better suited to the mobile computing environment. Their proposals are intended to cope with cost performance issues in data replication due to user mobility in mobile computing environment. To investigate their effectiveness, we provide access cost analysis and comparison on these strategies and the static replica allocation (SRA) strategy. We show that our proposed strategies outperform the SRA strategy when user mobility (inter-cell movement) is relatively low as compared with data access rate.

In this paper, a dynamic constructive algorithm for fault tolerant feedforward neural network, called DCFTA, is proposed. The algorithm starts with a network with single hidden neuron, and a new hidden unit is added dynamically to the network whenever it fails to converge. Before inserting the new hidden neuron into the network, only the weights connecting the new hidden neuron to the other neurons are trained (i. e. , updated) until there is no significant reduction of the output error. To generate a fault tolerant network, the relevance of each synaptic weight is estimated in each cycle, and only the weights which have their relevance less than a specified threshold are updated in that cycle. The loss of a connections between neurons (which are equivalent to stuck-at-0 faults) are assumed. The simulation results indicate that the network constructed by DCFTA has a significant fault tolerance ability.

Parallelism on heterogeneous machines brings cost effectiveness, but also raises a new set of complex and challenging problems. This paper addresses the problem of estimating the minimum time taken to execute a program on a fine-grained parallel machine composed of different types of processors. In an earlier publication, we took the first step in this direction by presenting a graph-construction method which partitions a given program into several homogeneous parts and incorporates timing constraints due to heterogeneous parallelism into each part. In this paper, to make the method easier to be applied in a scheduling framework and to demonstrate its practical utility, we present an efficient implementation method and compare the results of its use to the optimal schedule lengths obtained by enumerating all possible solutions. Experimental results for several different machine models indicate that this method can be effectively used to estimate a program's minimum execution time.

The basic operation characteristics of an asymmetric turnstile which transfers each electron one by one in one direction is described. A novel single electron counter circuit consisting of the asymmetric turnstiles, a load capacitor and an inverter which counts the number of high inputs is proposed. Monte Carlo circuit simulations reveal that the gate clock time of the counter circuit should be long enough to achieve allowable minimum error rate. The counter circuit implementing asymmetric single electron turnstiles is demonstrated to be applicable to a noise reduction system, a Winner-Take-All circuit and an artificial neuron circuit.

We present a new scheduling policy named Value-based Processor Allocation (VPA-k) for scheduling value-based transactions in a multiprocessor real-time database system. The value of a transaction represents the profit the transaction contributes to the system if it is completed before its deadline. Using VPA-k policy, the transactions with higher values are given higher priorities to execute first, while at most k percentage of the total processors are allocated to the urgent transactions dynamically. Through simulation experiments, VPA-k policy is shown to outperform other scheduling policies substantially in both maximizing the totally obtained values and minimizing the number of missed transactions.

This paper presents a new self-organization classification algorithm for remote-sensing images. Kohonen and other scholars have proposed self-organization algorithms. Kohonen's model easily converges to the local minimum by tuning the elaborate parameters. In addition to others, S. C. Amatur and Y. Takefuji have also proposed self-organization algorithm model. In their algorithm, the maximum neuron model (winner-take-all neuron model) is used where the parameter-tuning is not needed. The algorithm is able to shorten the computation time without a burden on the parameter-tuning. However, their model has a tendency to converge to the local minimum easily. To remove these obstacles produced by the two algorithms, we have proposed a new self-organization algorithm where these two algorithms are fused such that the advantages of the two algorithms are combined. The number of required neurons is the number of pixels multiplied by the number of clusters. The algorithm is composed of two stages: in the first stage we use the maximum self-organization algorithm until the state of the system converges to the local-minimum, then, the Kohonen self-organization algorithm is used in the last stage in order to improve the solution quality by escaping from the local minimum of the first stage. We have simulated a LANDSAT-TM image data with 500 pixel 100 pixel image and 8-bit gray scaled. The results justifies all our claims to the proposed algorithm.

In this paper, we propose an algorithm of classification by feature partitioning (CFP) which learns concepts in the batch mode. The proposed algorithm achieved almost the same predictive accuracies as the best results of a CFP algorithm presented by Guvenir and Sirin. However, our algorithm is not affected by parameters and the order of examples.

This paper proposes a new associative memory architecture using stochastic behavior in single electron tunneling (SET) devices. This memory stochastically extracts the pattern most similar to the input key pattern from the stored patterns in two matching modes: the voltage-domain matching mode and the time-domain one. In the former matching mode, ordinary associative memory operation can be performed. In the latter matching mode, a purely stochastic search can be performed. Even in this case, by repeating numerous searching trials, the order of similarity can be obtained. We propose a circuit using SET devices based on this architecture and demonstrate its basic operation with a simulation. By feeding the output pattern back to the input, this memory retrieves slightly dissimilar patterns consecutively. This function may be the key to developing highly intelligent information processing systems close to the human brain.

Gate performance for observing Coulomb oscillations and Coulomb diamonds are compared for two types of gated sub-µm double-barrier heterostructures. The first type of device contains modulation-doped barriers, whereas the second type of device contains a narrower band gap material for the well and no barriers with doped impurities. Both the Coulomb oscillations and Coulomb diamonds are modified irregularly as a function of gate voltage in the first type of device, while in the second type of device they are only systematically modified, reflecting atom-like properties of a quantum dot. This difference is explained in terms of the existence of impurities in the first type of device, which inhomogeneously deform the rotational symmetry of the lateral confining potential as the gate voltage is varied. The absence of impurities is the reason why we observe the atom-like properties only in the second type of device.

We propose a novel opto-electronic memory device using a single quantum dot (QD) and a logic device using coupled QDs (CQD) which performs (N)AND and (N)OR operations simultaneously. In both devices, occupation/unoccupation by a single electron in a QD is viewed as a bit 1/0 and data input/output (I/O) is performed by irradiation/absorption of photons. The (N)AND/(N)OR operations are performed by the relaxation of the electronic system to the Fock ground state which depends on the number of electrons in the CQD. When the device is constructed of semiconductor nanostructures, the main relaxation process is LA-phonon emission from an electron. Theoretical analysis of the device shows that (i) the error probability in the final state converges with the probability with which the system takes excited states at thermal equilibrium, i. e. , depends only on the dissipation energy and becomes smaller as the dissipation energy becomes larger, and (ii) the speed of operation depends on both the dissipation energy and dissipative interactions and becomes slower as the dissipation energy becomes larger if LA-phonon emission is taken into account. If the QDs are InAs cubes with sides of 10 nm and they are separated by the AlSb barrier with a width of 10 nm, the speed of operation and the error probability are estimated to be about 1 ns and about 0. 2 at 77 K, respectively. The basic idea of the device is applicable to two-dimensional (2D) pattern processing if the devices are arranged in a 2D array.

This letter gives a study of additionY=X+K mod 2w which is used in some cryptosystems as RC5. Our results enables us to express the differential and linear probability of addition as a function of addendK. To detect a good differential characteristics or linear approximation of a cryptosystem in which extended key is used as addend, we need to consider how the characteristics or approximations behave depending upon the value of the addend, which are clarified by our results.

This paper presents two types of two-dimensional (2-D) adaptive beamforming algorithm which have high rate of convergence. One is a linearly constrained minimum variance (LCMV) beamforming algorithm which minimizes the average output power of a beamformer, and the other is a generalized sidelobe canceler (GSC) algorithm which generalizes the notion of a linear constraint by using the multiple linear constraints. In both algorithms, we apply a 2-D lattice filter to an adaptive filtering since the 2-D lattice filter provides excellent properties compared to a transversal filter. In order to evaluate the validity of the algorithm, we perform computer simulations. The experimental results show that the algorithm can reject interference signals while maintaining the direction of desired signal, and can improve convergent performance.

In this paper, a new nonlinear integrator with positive phase shifts is proposed. Results of the digital simulation show that the nonlinear integrator has a better performance than the conventional one in a control system.

Though causal order of message delivery simplifies the design and development of distributed applications, the overhead of enforcing it is not negligible. We claim that a causal order algorithm which does not send any redundant information is efficient in the sense of communication overhead. We characterize and classify the redundant information into four categories: information regarding just delivered, already delivered, just replaced, and already replaced messages. We propose an efficient causal multicast algorithm which prevents propagation of these redundant information. Our algorithm sends less amount of control information needed to ensure causal order than other existing algorithms and can also be applied to systems whose communication channels are not FIFO. Since our algorithm's communication overhead increases relatively slowly as the number of processes increases, it shows good scalability feature. The potential of our algorithm is shown by simulation study.

In this paper, a design concept that offers ATM-VP connections with different protection levels is presented. The users have the choice to select the protection level they wish, the network transport service they need, and the worst cell loss they can tolerate at call set up time, and pay accordingly. Besides, an advanced adaptive traffic control scheme that simplifies call and cell processing is also presented. Many important functions such as call admission, VC-bandwidth reservation, cell-level congestion control, etc. are efficiently performed at the boundary of the backbone network. In this way is given a suitable answer to the important question: "How can future telecommunication networks based on ATM provide services with customized availability ?" A platform that outlines the potential interaction between restoration methods and congestion avoidance schemes is also obtained.

Detection of facial emotions are mainly addressed by computer vision researchers based on facial display. Also detection of vocal expressions of emotions is found in research work done by acoustic researchers. Most of these research paradigms are devoted purely to visual or purely to auditory human emotion detection. However we found that it is very interesting to consider both of these auditory and visual informations together, for processing, since we hope this kind of multimodal information processing will become a datum of information processing in future multimedia era. By several intensive subjective evaluation studies we found that human beings recognize Anger, happiness, Surprise and Dislike by their visual appearance, compared to voice only detection. When the audio track of each emotion clip is dubbed with a different type of auditory emotional expression, still Anger, Happiness and Surprise were video dominant. However Dislike emotion gave mixed responses to different speakers. In both studies we found that Sadness and Fear emotions were audio dominant. As a conclusion to the paper we propose a method of facial emotion detection by using a hybrid approach, which uses multimodal informations for facial emotion recognition.

Considering the pattern classification/recognition tasks, the influence of the activation function on fault tolerance property of feedforward neural networks is empirically investigated. The simulation results show that the activation function largely influences the fault tolerance and the generalization property of neural networks. It is found that, neural networks with symmetric sigmoid activation function are largely fault tolerant than the networks with asymmetric sigmoid function. However the close relation between the fault tolerance and the generalization property was not observed and the networks with asymmetric activation function slightly generalize better than the networks with the symmetric activation function. First, the influence of the activation function on fault tolerance property of neural networks is investigated on the XOR problem, then the results are generalized by evaluating the fault tolerance property of different NNs implementing different benchmark problems.

A deterministic pushdown automaton (dpda) having just one stack symbol is called a deterministic restricted one-counter automaton (droca). A deterministic one-counter automaton (doca) is a dpda having only one stack symbol, with the exception of a bottom-of-stack marker. The class of languages accepted by droca's which accept by final state is a proper subclass of the class of languages accepted by doca's. Valiant has proved the decidability of the equivalence problem for doca's and the undecidability of the inclusion problem for doca's. Thus the decidability of the equivalence problem for droca's is obvious. In this paper, we evaluate the upper bound of the length of the shortest input string (shortest witness) that disproves the inclusion for a pair of real-time droca's which accept by accept mode, and present a direct branching algorithm for checking the inclusion for a pair of languages accepted by these droca's. Then we show that the worst-case time complexity of our algorithm is polynomial in the size of these droca's.

Recently reported multidimensional geometrically uniform signal constellations (L MPSK and Decomposed-Lattice constellations) are joined in the term of Generalized Permutation Alphabets (GPA). Possibility of a binary isometric labeling of GPA's is completely characterized. An algorithm for constructing generating groups of PSK-type GPA is proposed. We show that this concept, when is extended to the lattice, gives rise to a class of new coset codes which perform out best codes listed in [11].