We present a method that uses optical flow to estimate facial muscle actions which can then be recognized as facial expressions. Facial expressions are the result of facial muscle actions which are triggered by the nerve impulses generated by emotions. The muscle actions cause the movement and deformation of facial skin and facial features such as eyes, mouth and nose. Since facial skin has the texture of a fine-grained organ, which helps in extracting the optical flow, we can extract muscle actions from external appearance. We are thus able to construct a facial expression recognition system based on optical flow data. We investigate the recogniton method in two ways. First, the optical-flow fields of skin movement is evaluated in muscle winsows, each of which defines one primary direction of muscle contraction to correctly extract muscle movement. Second, a fifteen dimensional feature vector is used to represent the most active points in terms of the flow variance through time and local spatial areas. The expression recognition system uses the feature vector to categorize the image sequences into several classes of facial expression. Preliminary experiments indicate an accuracy of approximately 80% when recognizing four types expressions: happiness, anger, disgust, and surprise.

The authors propose a new circular polarization method for a slot-coupled microstrip antenna. The simple configuration antenna in which a parasitically excited slot exists in addition to a directly excited slot generates circular polarization, by the effect of higher order modes. Experimental results and numerical results using the Spatial Network Method are described to verify the performance and discuss the circular polarization generating mechanism.

In this paper, two methods for improving the performance of DSBC (Dynamic bit allocation Sub-Band Coder) are proposed. One is to modify the bit allocation algorithm since conventional dynamic bit allocation algorithm is computationally heavy, while the other one is to weight the subband to reduce hissing noise. Instead of inserting the Gaussian noise in regeneration of empty band, the subband weighting method is to utilize shaped subband signal as empty band signal. The simulation result shows that computational complexity is reduced about 10%, and perceptual hissing noise is suppressed noticeably.

This letter presents a cancellation technique of parasitic poles of operational amplifier (op amp) in active filter design. To minimize the effect of the parasitic poles, a three-pole model of op amp is utilized. A second order highpass filter is evaluated both theoretically and numerically.

A reachability theorem for a class of live and safe free choice nets (LSFC nets) is presented in this paper. A necessary and sufficient condition for the reachability of a class of LSFC net is derived based on the associated digraph.

One of the basic problems in Information Theory, that is, the determination of the reliability function of binary symmetric channel, is studied by establishing the exponent of cardinality of intersection of two Hamming spheres.

This paper describes the frequency-domain analysis of a cross polarization interference canceller (XPIC) combined with a waveform equalizer. Analytical formula of the XPIC constructed by a multi-tap transversal filter is described under two-ray fading model. The XPIC performance relative to the correlation between the vertical and horizontal polarizations is also investigated by calculating the post-XPIC residual interference as a parameter of tap numbers. The improvement of disired-to-undesired signal ratio by the XPIC is discussed among three typical fading patterns. When co-polarization is dispersive with the same notch frequency (high correlation) and the cross polarization interference is non-dispersive, the XPIC performance is most degraded. The XPIC performance is better for fading events of non-dispersive co-polarization and dispersive cross polarization interference. Moreover, this paper shows the difference of the XPIC performance between the two types of receiver which obtain the reference signal from either post-equalizer or pre-equalizer.

FILT (Fast Inversion of Laplace Transform) method combined with MPMM (Modified Point Matching Method) is used to solve the problem of electromagnetic wave scattering by a circular cylinder with longitudinal slots. Some numerical results for the far field back scattering transient response are presented and discussed by comparing with the response for the perfect cylinder. The physical meaning of the transient wave form is discussed in detail. Also the precision of the analysis is checked carefully.

This paper describes the implementation types of private networks from the viewpoint of equipment and users, and utilization trends from low to high-speed services. Technology trends of private networks from the viewpoint of high-speed and broadband communication, multimedia communication, media access methods and intelligent communication are also examined. A concept for implementing future private networks in the Broadband ISDN (B-ISDN) age, and the major technology and the future efforts to achieve this implementation are discussed. Asynchronous Transfer Mode (ATM) is shown to be the key technology for integrating both various and heavy traffic media communication.

This paper describes theoretical analyses of a new CRV (Code Rule Violation) collision detection method applicable to an optical CSMA/CD LANs. The new method is based on the ideas combining a partial response (1,0, --1) circuit with the basic CRV collision detection method; the ideas are proposed in IEEE 802.3 committee. The circuit configuration of the new method, however, is different from that proposed in IEEE 802.3 committee. First, the paper outlines the basic CRV collision detection method and shows the problem that it can't detect a collision when identical bits are continuously received. Secondly, a practical circuit configuration of the new method is introduced. To realize large optical dynamic range and stability in collision detection, the circuit is designed not to detect CRVs on noisy zero level signals caused by signal negation with partial response circuits. The calculation formulae are then derived for the basic and the new method to determine the CRV rate with and without signal collision. At a bit rate of 10Mb/s, the CRV rate with and without signal collision is then calculated for the basic and the new method. Finally, the calculation results are compared with each other, and it is clarified that the new method is an effective method to increase CRV rate when signal collision occurs.

This letter presents a method of optimal design of IIR digital filters with the minimum number of interpolation points in magnitude response. Those filters are classified into four types and each of them is designed by a simple formula. Maximal flatness, ripple property, filter order, and steepness in transition band are discussed.

This paper presents a useful numerical approach based on a self-consistent finite-element method for solving stationary properties of third-order nonlinear guidedwave phenomena in a planar optical waveguide which supports nonlinearly coupled transverse-electric and transverse-magnetic modes. This method can be useful for the stability analysis as well by tracing intermediate solutions generated through iterative processes. Depending on the transitional behavior of the intermediate solutions we can identify the nonlinear excitation under consideration to be absolutely stable, quasi-stable, or unstable.

Corporate information communication systems are recently orienting the integration of various kinds of communication services. A PBX and a LAN have been indispensable in the realization of these systems. Some PBXs are orienting towards distributed switching architecture, and LAN systems are increasing their scale. Both requirements can be realized by means of a high speed loop optical fiber network. An integrated LAN should satisfy the required performance as a multimedia LAN and should be provided self-healing functions which heal failures automatically, to enable continuous operation of the entire system. This paper introduces a multimedia LAN, which attaches importance to the self-healing functions, called SHM-LAN. Its system architecture and its characteristics, such as efficient channel allocation are presented. This paper also deals with the self-healing functions and some applications.

The Fourier series expansion method is applied to the problem of EM waves guided by a dielectric grating with periodic surface-relief. The periodic structures of dielectric gratings are divided into two types: (1) surface-relief type and (2) modulated refractive index type. For the modulated dielectric grating, the rigorous solution is obtained using the space-harmonics expansion method, but the usual method cannot be applied to the grating whose dielectric constant distribution has discontinuities. Previously, we pointed out this issue and gave an improved method which can be applied to those problems. This method can also be applied to the surface-relief grating by approximating it as an assembly of stratified thin modulated dielectric gratings. In this paper, numerical analysis is given for gratings with sinusoidal surface and triangular surface for both perpendicular polarization (TE waves) and parallel polarization (TM waves). The validity of our method has been checked by increasing the number of modal expansion terms and the number of multilayers. The results are compared with those obtained by other numerical methods. The differences in the propagation characteristics between the surface-relief grating and the modulated dielectric grating in Bragg region are also discussed.

Recent information theoretical topics in cryptology are surveyed. Coding theorems are reviewed for Shannon's cipher system, Simmons' theory of authenticity, the wiretap channel, the secret sharing communication system, etc.

We propose a new multicast ATM switching network, called the Cascade Clos Broadcast Switching Network (CCB-SN). It has the following features. Firstly, the CCB-SN is robust not only for unicast calls but also for multicast calls. Therefore, deterioration of traffic handling performance due to a non-uniform traffic pattern does not occur. Secondly, in the CCB-SN, the cells are replicated after they are partly routed. And so, the delay time caused by the increment of replicas of cells is kept to a minimum.

For providing person-to-person communication as well as improving office work efficiency, demands for multimedia communication services have increased. This paper discusses Integrated Voice/Data LAN (IVDLAN) for handling multimedia communications. First, the IEEE 802.9 IVDLAN working group activities on a user access network which dedicatedly provides a multi-megabit service over an existing unshielded twisted telephone cabling system are summarized. Next, this paper reveals the first commercial IVDLAN product APEX 8000/10 which meets the IEEE 802.9 architecture. A set of an ISDN terminal and an IEEE 802.3 terminal are connected to the node through a terminal adapter. The line rate is 4.096 Mbit/s and the maximum line length is 600m. The IVDLAN node has an 128 Mbit/s burst switching module and a 64 Mbit/s multiple-rate circuit switching module. The flow control in a user-network interface and backpressure control in the burst switching module achieve the data communication with low packet loss.

A simple and practically accurate method for studying rectangular dielectric waveguides is proposed. This method can provide good evaluations of propagation constant and field distribution of the waveguides by analyzing improved equivalent guide models. In this paper, it is applied in studying a novel taper-formed waveguides system. The waveguides system was experimentally demonstrated of being advantageous in its lightwave transmission characteristics. Experimentally obtained results exhibited good agreement with the performances predicted by the system numerical analyses using the proposed waveguide study technique.

An arrangement technique of Gray-code table for signal constellations of modified QAM and Triangular-Shaped Signal Set is proposed. This is an optimum mapping technique for both two dimensional rectangular and hexagonal lattice-based constellations which cannot adapt the conventional Gray-code mapping. As our technique achieves most of neighboring symbols differ by one bit, the bit-error-probability (BEP) of the system becomes nearly equal to the symbol-error-probability (SEP). Furthermore, in consideration of this technique, we can have a constellation by which all neighboring symbols differ only by one bit for the rectangular lattice based constellation, but error-performances are worse than those of the original constellations because of a decrease of the corresponding decision regions.

A major problem in Doppler ultrasound has been the lack of quantitation of the magnitude of the blood velocity vector, particularly in intracardiac flow where the flow direction can't be estimated. This paper derives a new flow-mapping system for visualizing the two-dimensional blood velocity vector. The resultant velocity vector is the projection of the three-dimensional velocity vector on the beam scanning plane.