Introducing a mathematical model of image noise, we formalize the problem of fitting a conic to point data as statistical estimation. It is shown that the reliability of the fitted conic can be evaluated quantitatively in the form of the covariance tensor. We present a numerical scheme called renormalization for computing an optimal fit and at the same time evaluating its reliability. We also present a scheme for visualizing the reliability of the fit by means of the primary deviation pair. Our method is illustrated by showing simulations and real-image examples.

We study Lempel-Ziv-Yokoo algorithm [1, Algorithm 4] for universal data compression. In this paper, we give a simpler implementation of Lempel-Ziv-Yokoo algorithm than the original one [1, Algorithm 4] and show its asymptotic optimality for a stationary ergodic source.

We present a minimal lattice realization of MIMO linear discrete-time systems which interpolate the desired Markov and covariance parameters. The minimal lattice realization is derived via a recursive construction algorithm based on the state space description and it parametrizes all the interpolants.

For reducing requirement of large memory and minimizing computation complexity in a large-vocabulary continuous speech recognition system, speech segmentation plays an important role in speech recognition systems. In this paper, we formulate the speech segmentation as a two-phase problem. Phase 1 (frame labeling) involves labeling frames of speech data. Frames are classified into three types: (1) silence, (2) consonant and (3) vowel according to two segmentation features. In phase 2 (syllabic unit segmentation) we apply the concept of transition states to segment continuous speech data into syllabic units based on the labeled frames. The novel class of hyperrectangular composite neural networks (HRCNNs) is used to cluster frames. The HRCNNs integrate the rule-based approach and neural network paradigms, therefore, this special hybrid system may neutralize the disadvantages of each alternative. The parameters of the trained HRCNNs are utilized to extract both crisp and fuzzy classification rules. In our experiments, a database containing continuous reading-rate Mandarin speech recorded from newscast was utilized to illustrate the performance of the proposed speaker independent speech segmentation system. The effectiveness of the proposed segmentation system is confirmed by the experimental results.

This paper first presents the performance analysis of the NACF algorithm. The results show the possibility of the degradation in the convergence speed. To improve the convergence speed, the bias term is introduced into the NACF algorithm and its efficiency is investigated through the computer simulations.

In this paper, quantization method which can keep the phase and gain characteristics of a reference filter is proposed. The proposed method uses a genetic algorithm and a simulated annealing algorithm. The objective function used in this method is described with two kinds of weighting functions for identifying the phase and gain characteristics respectively. Therefore, the quantization accuracy on the gain characteristic is independent of the accuracy on the phase characteristic. Further, the proposed algorithm can be applied to any types of filters, because the chromosome expresses only their coefficients values. The efficiency of the proposed algorithm is verified by some experiments.

In this paper, DSPs, of which memory addresses are pointed by special purpose registers (address registers: ARs), are assumed, and methods to derive an efficient memory access pattern for those DSPs proposed. In such DSPs, programmers must take care for efficient allocation of memory space as well as effective use of registers, in order to derive an efficient program in the sense of execution period. In this paper, memory addresses and AR update operations are modeled by an access graph, and a novel memory allocation method is presented. This method removes cycles and forks in a given access graph, and decides an address location of variables in memory space with less overhead. In order to utileze multiple ARs, methods to assign variables into ARs are investigated. The proposed methods are applied to the compiler for DSP56000 and are proved to be effective by generated codes for several examples.

We propose the Kernel MUSIC algorithm as an improvement over the conventional MUSIC algorithm. This algorithm is based on the orthogonality between the image and kernel space of an Hermitian mapping constructed from the received data. Spatial smoothing, needed to apply the MUSIC algorithm to coherent signals, is interpreted as constructing procedure of the Hermitian mapping into the subspace spanned by the constituent vectors of the received data. We also propose a new spatial smoothing technique which can remove the redundancy included in the image space of the mapping and discuss that the removal of redundancy is essential for improvement of resolution. By computer simulation, we show advantages of the Kernel MUSIC algorithm over the conventional one, that is, the reduction of processing time and improvement of resolution. Finally, we apply the Kernel MUSIC algorithm to the Laser Microvision, an optical misroscope we are developing, and verify that this algorithm has about two times higher resolution than that of the Fourier transform method.

This paper proposes a fast timing recovery method with a decision feedback equalizer for baudrate sampling. The proposed method features two special techniques. The first one is for coarse estimation of the sampling phase. Internal signals of the oversampled analog-to-digital converter at different phases are directly taken out for parallel evaluation. The second technique provides fine tuning with a phase-modification stepsize which is adaptively controlled by the residual intersymbol interference. Simulation results by a full-duplex digital transmission system with a multilevel line code show superiority of the proposed method. The coarse timing estimation and the fine tuning reduce 75% and 40% of the time required by the conventional method,respectively. The overall saving in timing recovery is almost 60% over the conventional method. The proposed method could easily be extended to other applications with a decision feedback equalizer.

Internet is evolving to be an integrated service platform for promising future global information infrastructure. New multimedia services, e.g., Video-on-Demand, are expected to be provided on this platform, in conjuction with the existing rich information exchange and sharing services on Internet, such as E-mail, WWW. The paper proposes a service system architecture for Video-on -Demand on Internet, named MINS (Multimedia Internet Navigation System), which is capable of both realtime audio/video transmission and realtime and VCR like plya-out. MINS is composed of 1) servers maintaining both audio/video program lists and contents, and 2) clients retrieving the programs and play-out them. The paper focuses on the communication architecture in MINS, e.g., flow control, realtime media synchronization control, and audio/video session control. The performance of the prototype MINS is also evaluated in the paper.

The effect of silicone vapour concentration on the contact failure was examined by using micro relays and motor brush-slip ring(commutator) contacts, [(CH3) 2SiO]4: D4 was used as a vapour source of silicone contamination. Because the influence of the vapour of the silicone on the contact surface can not be avoided at all times due to its gradual evaporation in the atmosphere. The contact failure caused by the silicone vapour was confirmed as formation of SiO2 on the contact surfaceby analysis of EPMA and XPS. A minimum limiting concentration level which does not affect contact reliability was found. This limiting level was 10 ppm(O.13mg/l). Validity of the limiting level was confirmed by the relationships among concentration, temperature, SiO2 film thickness and contact resistance. Furthermore, the effect of the degree of silicone polymerization on the limiting concentration was derived by an empirical formula. This silicone is found to have polymerization degree larger than D7: n=7. These results were confirmed by the contact failure data due to the silicone contamination.

Adapting the principle of parametron oscillation, a small implantable temperature sensor requiring no internal power supply is described. Since this sensor's oscillation frequency is half that of the excitation frequency, the oscillated signal can be measured from the reception side, free of any signal, interference, simply by positioning the sensor and the excitation antenna so that; 1) they are separated up to 95 cm in the air; 2) a 41 cm gap, the phantom equivalent of the thickness of the human abdomen maintain between them. In the temperature-dependent quartz resonator sensor, oscillation occurs only when frequency and temperature correspond. The excitation power is then adjusted so that the frequency bandwidth narrows. As a result, the margin of error in measuring the temperature is minimized; (0.07).

Many activities are being promoted for the coming multimedia age. In this paper, background information for multimedia communications is followed by an outline of joint tests in multimedia communications with some examples of the projects and applications. These trials are also explained from the aspects of project specifications, which include application classifications and details of multimedia-on-demand offerings, as well as technical issues in experimental environments which mainly include those related to ATM technology.

A self-organizing neural network model of spatio-temporal visual receptive fields is proposed. It consists of a one-layer linear learning network with multiple temporal input channels, and each temporal channel has different impulse response. Every weight of the learning network is modified according to a Hebb-type learning algorithm proposed by Sanger. It is shown by simulation studies that various types of spatio-temporal receptive fields are self-organized by the network with random noise inputs. Some of them have similar response characteristics to X- and Y-type cells found in mammalian retina. The properties of receptive fields obtained by the network are analyzed theoretically. It is shown that only circularly symmetric receptive fields change their spatio-temporal characteristics depending on the bias of inputs. In particular, when the inputs are non-zero mean, the temporal properties of center-surround type receptive fields become heterogeneous and alter depending on the positions in the receptive fields.

A main result in this paper is that for a Hopfield-type neural circuit with a symmetric connection matrix T, the negative semidenfiniteness of T is a necessary and sufficient condition for absolute exponential stability. While this result extends one of absolute stability in Forti, et al. [1], its proof given in this paper is simpler, which is completed by an approach different from one used in Forti et al. [1]. The most significant consequence is that the class of neural networks with negative semidefinite matrices T is the largest class of symmetric networks that can be employed for embedding and solving optimization problem with global exponential rate of convergence to the optimal solution and without the risk of spurious responses.

String grammars (languages) have been extensively studied from 60's. On the other hand, the transformational grammar, proposed by N. Chomsky, contains the transformation from the set of derivation trees of context-free language to the surface set. And the grammar regarded a tree as an input sentence to some transducer. After that from latter half of 60's, the studies of acceptors, transducers, and so on, whose input is a tree, have been studied extensively. And recently some pushdown tree automata were introduced, and their fundamental properties and some other various properties were investigated [11]-[17]. Furthermore, a top-down pushdown tree transducer (t-PDTT for short), which is an extension of a top-down pushdown automaton (t-PDTA for short), was introduced and its fundamental properties were investigated [19]. In this paper, we define the various subclasses of context-free tree grammar (CFTG for short) by the combination of variables contained in the rules. Furthermore, we consider a monadic case of CFTG which is a special case of CFTG. Based on these definitions, we classify the subclasses of CFTG, and we investigate some inclusion properties of subclasses of CFTL (where CFTL indicates the class of context-free tree languages).

In the usual optical flow detection, the gradient constraint, which expresses the relationship between the gradient of the image intensity and its motion, is combined with the least-squares criterion. This criterion means assuming that only the time derivative of the image intensity contains noise. In this paper, we assume that all image derivatives contain noise and derive a new optical flow detection technique. Since this method requires the knowledge about the covariance matrix of the noise, we also discuss a method for its estimation. Our experiments show that the proposed method can compute optical flow more accurately than the conventional method.

In future mobile radio, high-speed transmission and efficient spectrum utilization will be important. However, multipath propagation with large delay difference and cochannel interference are obstacles to the advanced mobile communication system. An adaptive antenna can suppress multipath signals and cochannel interference signals. This paper reviews basic performance of multipath fading reduction and cochannel interference suppression using the adaptive antenna. After a brief explanation of adaptive antenna concepts, we show simulation and experimental results of the fading reduction. It is pointed out that the adaptive antenna cancels multipath signals with large delay difference strongly. This feature is very important for high-speed TDMA systems. Moreover, it is shown from simulation results that the adaptive antenna improves the spectrum efficiency by suppressing the cochannel interference signals.

Exploring enormous state graphs represented implicitly by ordered binary decision diagrams (OBDDs) is one of the most successful applications of OBDDs. However, our worst-case analysis of implicit graph representations by OBDDs shows that there are cases where OBDD representations are not optimal and require more space than adjacency lists. As an improvement, we propose a new type of BDDs, called Patricia BDDs, which are capable of implicit representation of graphs while their worst-case sizes are kept equal or less than adjacency lists and OBDDs.

A blind equalization algorithm which makes use of the Stop" region of the Stop-and-Go algorithm is proposed. By adaptively updating the tap weights at the Stop region as well, it is intended to improve the convergence property of the Stop-and-Go algorthm. The performance of the proposed algorithm is compared with the conventional Stop-and-Go algorithm using various communication channels. Simulation results indicate the improvement of the convergence speed while maintaining or possibly lowering the residual error.