In the field of the restoration of motion blurred images, several deterministic techniques are proposed. Motion blurred images, however, include an ambiguity intrinsically. Therefore, the statistical restoration techniques are suited for the aim. In this paper, we describe the statistical properties of images blurred by various motions in mean square error sense since the discussions on the subject have been not presented in the past. The mean square errors of images blurred by motions and of images restored by Wiener filtering are formulated. The relations between the mean square errors and the extent of the motion blur or the correlation coefficient of the object are presented. The differences of the mean square errors among the various motions are discussed. Furthermore, these formulations are expanded to the motion blurred images imbedded by additive noise.
In database normalization theory, multivalued dependencies (MVDs) play an important and much-studied role. However, their semantic aspects in a conceptual schema have not been so thoroughly studied. This is one of the causes which prevent the theory from being applied to actual database design. This paper clarifies that, in a conceptual schema, MVDs have at least two different meanings. In this paper we first define a data model, called a functional dependency graph. We show that a relation scheme can be constructed from the functional dependency graph. Then, we show that MVDs which hold in the relation scheme correspond to certain substructures of the functional dependency graph. This is one meaning of MVDs. Another meaning of MVDs is that MVDs are embodied by unions of relations having different relation schemes. We show that a semantic problem with transitively specified MVDs does not arise when we interpret MVDs to have the latter meaning. It becomes easier to obtain MVDs from out real world by using the results just described. This helps us apply database normalization theory to actual database design.
Hideo MIYAHARA Toshiharu HASEGAWA
This paper considers the viability of a flexible multiplex structure for circuit- and packet-switched traffic for reducing and/or eliminate unused time intervals. These unused time intervals generally occur in the communication link despite traffic requiring transmission, when frame and packet lengths are restricted to a constant value, as is done in ordinary integrated multiplex structure where, when variable frame length is used to reduce these unused time intervals, synchronous assignment of time slots for circuit switched traffic becomes impossible. The method given here introduces two types of variable length packets, one for conveying circuit switched traffic only and the other for packet switching, to obtain approximately the same effect as when frame and packet are allowed variable length. The results of a performance evaluation of the method shows that this multiplexing structure will function.
The selection and scanning diversity effects on the error rate performance of digital FM with limiter discriminator detection in the fast Rayleigh fading signal environment are theoretically analyzed. For evaluating the average error rate with diversity, the simplified method of deriving the conditional error rate, including the effect of random FM noise, is proposed. The method is based on that the average error rate for no diversity and the conditional error rate are related to the Laplace transform in the Rayleigh fading signal environment. Then, the diversity effects are presented in the closed form expression for both symmetrical and asymmetrical fading power spectra.
A perspective projection is a useful two dimensional representation of a three dimensional world. In general, it is difficult to reconstruct the three dimensional information from the two dimensional image because some of the three dimensional information is lost in the perspective projection. The inverse of perspective projections is difined as a procedure to reconstruct the three dimensional information from a perspective projection. Characteristics of perspective projections, especially their degeneracy, are discussed in detail. Generally, even by treating several points at the same time and knowing the geometric relations between those points, the problem cannot be solved uniquely. We give several sufficient conditions for making the inverse unique. We discuss the inverse of perspective projections mainly from the theoretical point of view in this paper but possible applications of the method are also discussed briefly. A computer simulation of the algorithm is also given.
Takeshi AGUI Yukihiro ARAI Masayuki NAKAJIMA
An algebraic operating system for pixel chains and codes difined for binary figures is described. Several kinds of arithmetic four operations are difined for the pixel chains and codes, and properties of the arithmetic operations and binary figures are studied to clear the problems. The algebraic operating system adopted in this paper enables us to magnify and reduce binary figures. Several examples for such operations are illustrated. In appendix, an operating system satisfying the axioms of algebraic group is described.
Shoji SAKURAZAWA Yoshihide IGARASHI Yukio SHIBATA
In this paper we discuss some combinatorial problems related to the permutation network devised by Waksman et al. NUM(π) means the number of configurations of the permutation network producing a permutation π. The main result in this paper is that 23N-5-(log2N(log2N+1))/2 is a lower bound on the number of elements in the set {π
Tadatoshi TANIFUJI Tsuneo HORIGUCHI Masamitsu TOKUDA Kiyohide MIYAKE Tadashi FUKUZAWA
Far-end measurement instrument for baseband frequency response, based on swept frequency method, is designed and constructed. In constituting the measurement instrument, it is important to stabilize modulation frequency characteristics of laser diode which is significantly affected by externally reflected light to the laser diode. It has been found that by coupling laser diode to a fiber with obliquely polished end face, the influence of externally reflected light can be negligibly small. Optimum coupling conditions, concerning tilt angle of fiber end and distance between fiber and laser facet are clarified. Moreover, automatic power control and temperature control circuit has been applied to the above laser to fiber coupler and far-end measurement instrument of baseband frequency response has been constructed. As a result, sufficient stabilization of modulation frequency characteristics to measure baseband frequency response of graded-index optical fiber has been attained.
Chiaki HISHINUMA Masaichi KAJIWARA Masaki ITOH
Multiple communication Bus Network (MCNET) is proposed. This network interconnect, all exchanges in a specified area by one large scale transmission line to aim at network cost reduction and to secure adaptability for traffic variation.
In this letter we propose a multipurpose fiber optic sensor composed of a fiber, a spherical micro lens, and a sensing mirror. The analog sensor converts a physical quantity such as oscillation or displacement to a quantity of light. Replacing the lens with one of a different diameter makes it possible to use the new sensor several different ways, such as for sensing flow rate, pressure or temperature. Experimental results show that the dynamic range of the mirror's position displacement changes from 7.5 µm to 500 µm and that the dynamic range of the conversion gain changes about 20 dB when the lens is changed from 0.5 mm to 5.0 mm in diameter.
Heihachi SATO Hidenori HAYASHI Mitsugu SUGIYAMA Shuji TSUCHIYA
By analyzing reflectance variation of human blood involving various diseases under DC magnetic field at Ar+ 0.5 µm laser line, we can obtain new informations for diagnosis on whether the disease is malignant or benign.
A method of reducing reflection for a lattice soliton at the termination of a nonlinear LC line is investigated. It is shown that a termination with tapering resistances gives considerably less reflection than that of the characteristic impedance of the line of the small-signal limit.
A method of detecting and correcting digital errors in PCM-decoded speech signals using decoded signal information only, has been proposed by the author. This letter describes a new method to detect the errors using phase spectrum and an improvement in the spectral subtraction method used in a previous report.