This paper presents a scale invariant face detection and classification method which uses shift invariant features extracted from a Log-Polar image. Scale changes of a face in an image are represented as shift along the horizontal axis in the Log-Polar image. In order to obtain scale invariant features, shift invariant features are extracted from each row of the Log-Polar image. Autocorrelations, Fourier spectrum, and PARCOR coefficients are used as shift invariant features. These features are then combined with simple classification methods based on Linear Discriminant Analysis to realize scale invariant face detection and classification. The effectiveness of the proposed face detection method is confirmed by experiments using face images captured under different scales, backgrounds, illuminations, and dates. To evaluate the proposed face classification method, we performed experiments using 2,800 face images with 7 scales under 2 different backgrounds and face images of 52 persons.

In this paper, we present a dynamic output feedback controller design technique for robust decentralized stabilization of uncertain large-scale systems with time-delay in the subsystem interconnections. Based on Lyapunov second method, a sufficient condition for the stability, is derived in terms of three linear matrix inequalities (LMI). The solutions of the LMIs can be easily obtained using efficient convex optimization techniques. A numerical example is given to illustrate the proposed method.

To make clear numerically the scattering characteristics for a body embedded in a random medium, we need to analyze the bistatic cross-section (BCS). The scattering problem can be analyzed as a boundary value problem by using current generator method. The fourth moment of Green's functions in the random medium, which is necessary for the analysis, is obtained approximately by two-scale method. We analyze numerically the BCS of conducting circular cylinders in continuous random media, which are assumed to fluctuate about the dielectric constant of free space. The numerical results agree well with the law of energy conservation. The effects of random media on the BCS are also clarified numerically.

Although a number of car-traffic simulators have been developed for various purposes, none of the existing simulators enhance the simulation accuracy using sensor data or allow the system structure to re-configure the system structure depending on the application. Our goal was to develop a highly accurate, highly modular, flexible, and scalable micro-model car-traffic simulation system. The HLA (High Level Architecture) was applied to every system module as a standard interface between each module. This allows an efficient means for evaluating and validating a variety of micro-model simulation schemes. Our ongoing projects consist of running several identical simulations concurrently, with different parameter sets. By sending the results of these simulations to a manager module, which analyzes both the parameter sets and the simulated results, the manager module can evaluate the best-simulated results and determine the next action by comparing these results with the sensor data. In this system, the sensor data or the statistical data on the flow of traffic, obtained by monitoring real roads, is used to improve the simulation accuracy. Future systems are being planned to employ real time sensor data, where the input of the data occurs at almost real time speed. In this paper, we discuss the design of a HLA-based car-traffic simulation system and the construction of a sensor-data fusion algorithm. We also discuss our preliminary evaluation of the results obtained with this system. The results show that the proposed fusion algorithm can adjust the simulation accuracy to the logged sensor data within a difference of 5% (minimum 1.5%) in a specific time period. We also found that simulations with 500 different parameter sets can be executed within 5 minutes using 8 simulator modules.

This paper presents a Path Bandwidth Management (PBM) model for large-scale networks that leads to an almost optimal PB allocation, under constraints posed by the installed bandwidth in the transmission links of the network. The PB allocation procedure is driven from a traffic demand matrix and consists of three phases. In the first phase, a suitable decomposition of the whole network takes place, where the large-scale network is split to a set of one-level sub-networks. In the second phase, the optimization algorithm developed for one-level telecommunication networks is applied to each sub-network in order to define the optimal PB allocation. The criterion for optimization is to minimize the worst Call Blocking Probability (CBP) of all switching pairs of the sub-network. In the third phase, composition of the sub-networks takes place in a successive way, which leads to the final PB allocation of the large-scale network. As the large-scale network is built up from optimized sub-networks, an almost optimal PB allocation is anticipated. For evaluation, the worst resultant CBP of the proposed scheme is compared with that obtained by the optimal PB allocation procedure in order to prove its optimality and efficiency. We choose a set of large-scale networks whose size is not very large so that we can apply the optimization algorithm developed for one-level telecom networks for defining its optimal bandwidth allocation. Extensive evaluation of the PBM model has showed that the worst resultant CBP is about 2% above the optimal value, which is a satisfactory result. The proposed PBM scheme is explained by means of an application example.

We gave in [1] the software and hardware algorithms for reconfiguring 1 1/2-track switch 2-D mesh arrays with faults of processing elements, avoiding them. This paper shows an implementation of the hardware algorithm using an FPGA device, and by the logical simulation confirms the correctness of the behavior and evaluates reconfiguration time. From the result it is found that a self-repairable system is realizable and the system is useful for the run-time as well as fabrication-time reconfiguration because it requires no host computer to execute the reconfiguration algorithm and the reconfiguration time is very short.

We propose a rough classification system using Hierarchical Learning Vector Quantization (HLVQ) for large scale classification problems which involve many categories. HLVQ of proposed system divides categories hierarchically in the feature space, makes a tree and multiplies the nodes down the hierarchy. The feature space is divided by a few codebook vectors in each layer. The adjacent feature spaces overlap at the borders. HLVQ classification is both speedy and accurate due to the hierarchical architecture and the overlapping technique. In a classification experiment using ETL9B, the largest database of handwritten characters in Japan, (it contains a total of 607,200 samples from 3036 categories) the speed and accuracy of classification by HLVQ was found to be higher than that by Self-Organizing feature Map (SOM) and Learning Vector Quantization methods. We demonstrate that the classification rate of the proposed system which uses multi-codebook vectors for each category under HLVQ can achieve higher speed and accuracy than that of systems which use average vectors.

In this paper, we have proposed a new method for diffuse liver disease classification with sonogram, including the normal liver, hepatitis and cirrhosis, from a new point of view "scale. " The new system utilizes a multiscale analysis tool, called wavelet transforms, to analyze the ultrasonic liver images. A new set of features consisting of second order statistics derived from the wavelet transformed images is employed. From these features, we have found that the third scale is the representative scale for the classification of the considered liver diseases, and the horizontal wavelet transform can improve the representation of the corresponding features. Experimental results show that our method can achieve about 88% correct classification rate which is superior to other measures such as the co-occurrence matrices, the Fourier power spectrum, and the texture spectrum. This implies that our feature set can access the granularity from sonogram more effectively. It should be pointed out that our features are powerful for discriminating the normal livers from the cirrhosis because there is no misclassification samples between the normal liver and the cirrhosis sets. In addition, the experimental results also verify the usefulness of "scale" because our multiscale feature set can gain eighteen percent advantage over the direct use of the statistical features. This means that the wavelet transform at proper scales can effectively increase the distances among the statistical feature clusters of different liver diseases.

Standard client-server workflow management systems have an intrinsic scalability limitation, the central server, which represents a bottleneck for large-scale applications. This server is also a single failure point that may disable the whole system. We propose a fully distributed architecture for workflow management systems. It is based on the idea that the case (an instance of the process) migrates from host to host, following a process plan, while the case activities are executed. This basic architecture is improved so that other requirements for Workflow Management Systems, besides scalability, are also contemplated. A CORBA-based implementation of such architecture is discussed, with its limitations, advantages and project decisions described.

A large-scale primarily public system requires in addition to high reliability, a broad range of applications from control to information services. As construction is phased-in this system must be flexible, changeable and able to grows as the needs arise. However, a changing a system may lead to loss of reliability. A system that is able to change and grow in a reliable and stable manner is called an assurance system and the technology it uses is called assurance technology. This paper describes the basic technology, phased-in system construction and so on of assurance technology based on an autonomous decentralized system. It further discusses application of assurance technology to ATOS as an example of a large-scale transport operation control system. Note: ATOS; Autonomous Decentralized Transport Operation Control System

In the phase locked loop (PLL) frequency synthesizer which is used in a higher frequency region, the prescaler method is employed in order to increase the operating frequency of the programmable divider. However, since the fixed divider whose division ratio is same as the prescaler is installed at the following stage of the reference divider, the reference frequency is decreased and the performance of the PLL frequency synthesizer is degraded. The prescaler PLL frequency synthesizer using multi-programmable divider is one of the counter measures answering the request. In this paper we propose the reduction of the number of programmable dividers by using the (N+1/2) programmable divider. The effectiveness of the proposed method is confirmed by experimental results.

This paper describes a distributed traffic control scheme for large multi-stage ATM switching systems. When a new virtual circuit is to be added from some source line-interface unit (LU) to a destination LU, the system must find an optimal path through the system to accommodate the new circuit. Conventional systems have a central control processor and control lines to manage the bandwidth of all the links in the systems. The central control processor handles all the virtual circuits, but have trouble doing this when the switching system becomes large because of the limited ability of the central processor to handle the number of virtual circuits. A large switching system with Tbit/s-class throughput requires a distributed traffic control scheme. In our proposed switching system, each port of the basic switches has its own traffic monitor. Operation, administration, and maintenance (OAM) cells that are defined inside the system carry the path-congestion information to the LUs, enabling each LU to route new virtual circuits independently. A central control processor and control lines are not required. The performance of the proposed system depends on the interval between OAM cells. This paper shows how an optimal interval can be determined in order to maximize the bandwidth for user cells. This traffic control scheme will suit future Tbit/s ATM switching systems.

Fuzzy rule-based edge detection using multiscale edge images is proposed. In this method, the edge image is obtained by fuzzy approximate reasoning from multiscale edge images which are obtained by derivative operators with various window sizes. The effect of utilizing multiscale edge images for edge detection is already known, but how to design the rules for deciding edges from multiscale edge images is not clarified yet. In this paper, the rules are represented in a fuzzy style, since edges are usually defined ambiguously, and the fuzzy rules are designed optimally by a training method. Here, the fuzzy approximate reasoning is expressed as a nonlinear function of the multiscale edge image data, and the nonlinear function is optimized so that the mean square error of the edge detection be the minimum. Computer simulations verify its high performance for actual images.

This paper presents a robust and nonblocking group membership protocol for large-scale distributed systems. This protocol uses the causal relation between membership-updating messages (i. e. , those specifying the adding and deleting of members) and allows the messages to be executed in a nonblocking manner. It differs from conventional group membership protocols in the following points: (1) neither global locking nor global synchronization is required; (2) membership-updating messages can be issued without being synchronized with each other, and they can be executed immediately after their arrival. The proposed protocol therefore is highly scalable, and is more tolerant to node and network failures and to network partitions than are the conventional protocols. This paper proves that the proposed protocol works properly as long as messages can eventually be received by their destinations. This paper also discusses some design issues, such as multicast communication of the regular messages, fault tolerance and application to reliable communication protocols (e. g. , TCP/IP).

As VLSI technology has developed, the interest in implementing an entire or significant part of a parallel computer system using wafer scale integration is growing. The major problem for the case is the possibility of drastically low yield and/or reliability of the system if there is no strategy for coping with such situations. Various strategies to restructure the faulty physical system into the fault-free target logical system are described in the literature [1]-[5]. In this paper, we propose an efficient approximate method which can reconstruct the 1 1/2 track-switch mesh arrays with faulty PEs using hardware as well as software. A logical circuit added to each PE and a network connecting the circuits are used to decide spare PEs which compensate for faulty PEs. The hardware compexity of each circuit is much less than that of a PE where the size of each additional circuit is independent of array sizes and constant. By using the exclusive hardware scheme, a built-in self-reconfigurable system without using a host computer is realizable and the time for reconfiguring arrays becomes very short. The simulation result of the performance of the method shows that the reconstructing efficiency of our algorithm is a little less than those of the exaustive and Shigei's ones [6] and [7], but much better than that of the neural one [3]. We also compare the time complexities of reconstructions by hardware as well as software, and the hardware complexity in terms of the number of gates in the logical circuit added to each PE among the other methods.

This paper proposes implementing guaranteed frame rate (GFR) service using the available bit rate (ABR) control mechanism in large-scale networks. GFR is being standardized as a new ATM service category to provide a minimum cell rate (MCR) guarantee to each virtual channel (VC) at the frame level. Although ABR also can support MCR, a source must adjust its cell emission rate according to the network congestion indication. In contrast, GFR service is intended for users who are not equipped to comply with the source behavior rules required by ABR. It is expected that many existing users will fall into this category. As one implementation of GFR, weighted round robin (WRR) with per-VC queueing at each switch is well known. However, WRR is hard to implement in a switch supporting a large number of VCs because it needs to determine in one cell time which VC queue should be served. In addition, it may result in ineffective bandwidth utilization at the network level because its control mechanism is closed at the node level. On the other hand, progress in ABR service standardization has led to the development of some ABR control algorithms that can handle a large number of connections. Thus, we propose implementing GFR using an already developed ABR control mechanism that can cope with many connections. It consists of an explicit rate (ER) control mechanism and a virtual source/virtual destination (VS/VD) mechanism. Allocating VSs/VDs to edge switches and ER control to backbone switches enables us to apply ABR control up to the entrance of a network, which results in effective bandwidth utilization at the network level. Our method also makes it possible to share resources between GFR and ABR connections, which decreases the link cost. Through simulation analysis, we show that our method can work better than WRR under various traffic conditions.

A new feature parameter based on a discrete wavelet transform is proposed for word boundary detection of isolated utterances. The sum of standard deviation of wavelet coefficients in the third coarse and weighted first detailed scale is defined as a new feature parameter for endpoint detection. Experimental results demonstrate the superiority of the proposed feature to the conventional ones in capturing word boundaries even in noisy speech.

A content-based image retrieval scheme based on scale-space theory is proposed. Instead of considering all scales for image retrieval, the proposed algorithm utilizes a modified histogram intersection method to compute the relative scale between a query image and a candidate image. The proposed method has been applied to various images and the performance improvement has been verified.

This paper outlines the modeling requirements of integrated circuit (IC) fabrication processes that have lead to and sustained the development of computer-aided design of technology (i. e. TCAD). Over a period spanning more than two decades the importance of TCAD modeling and the complexity of required models has grown steadily. The paper also illustrates typical applications where TCAD has been powerful and strategic to IC scaling of processes. Finally, the future issues of atomic-scale modeling and the need for an hierarchical approach to capture and use such detailed information at higher levels of simulation are discussed.

This paper describes a flip-flop circuit using a directly controlled emitter-follower with a diode-feedback level stabilizer (DC-DF) and a resistor-feedback level stabilizer (DC-RF) for low-power multi-GHz prescalers. The new flip-flop circuit reduces the emitter-follower current and gains both high-frequency operation and low-power. A dual modulus (4/5) prescaler using this circuit technology was fabricated with a 0.35 µm BiCMOS process. The current draw of the prescaler using the DC-RF is 34% smaller than conventional LCML circuits. The DC-RF prescaler operates at 2.11 GHz with a total current consumption of 1.03 mA. In addition, the circuit operates with a supply voltage of down to 2.4 V by using the resistor level-shift clock-driver.