In computer vision, the interpretation of 3D motion of an object in the physical world is an important task. This study proposes a 3D motion interpretation method which uses a neural network system consisting of three kinds of neural networks. This system estimates the solutions of 3D motion of an object by interpreting three optical flow (OF-motion vector field calculated from images) patterns obtained at the different view points for the same object. In the system, OF normalization network is used to normalize diverse OF patterns into the normalized OF format. Then 2D motion interpretation network is used to interpret the normalized OF pattern and to obtain the object's projected motion onto an image plane. Finally, 3D motion interpretation network totally interprets the three sets of the projected motions and it derives the solutions of the object's 3D motion from the inputs. A complex numbered version of the back-propagation (Complex-BP) algorithm is applied to OF normalization netwerk and to 2D motion interpretation network, so that these networks can learn graphical patterns as complex numbers. Also a 3D vector version of the back-propagation (3DV-BP) algorithm is applied to 3D motion interpretation network so that the network can learn the spatial relationship between the object's 3D motion and the corresponding three OF patterns. Though the interpretation system is trained for only basic 3D motions consisting of a single motion component, the system can interpret unknown multiple 3D motions consisting of several motion components. The generalization capacity of the proposed system was confirmed using diverse test patterns. Also the robustness of the system to noise was probed experimentally. The experimental results showed that this method has suitable features for applying to real images.

Learning correctly from queries" is a formal learning model proposed by Angluin. In this model, for a class Γ of language representations, a learner asks queries to a teacher of an unknown language Lq which can be represented by some GqΓ, and eventually outputs a language representation GΓ which represents Lq and halts. An algorithm (leaner) A is said to learn a class of languages represented by Γ in the weak definition if the time complexity of A is some polynomial of n and m, where n is the minimum size of the lagunage representations in Γ which represent Lq, and m is the maximum length of the counterexamples returned in an execution. On the other band, A is said to learn represented by Γ in the strong definition if at any point τ of the execution, the time consumed up to τ is some polynomial of n and m, where n is the same as above, and m is the maximum length of the counterexamples returned up to τ. In this paper, adequacy of the model is examined, and it is shown that both in the weak and strong definitions, there exist learners which extract a long counterexample, and identify Lq by using equivalence queries exhaustively. For example, there exists a learner which learns the class CFL of context-free languages represented by the class CFG of context-free grammars in the weak definition using only equivalence queries. Next, two restrictions concerning with learnability criteria are introduced. Proper termination condition is that when a teacher replies with yes" to an equivalence query, then the learner must halt immediately. The other condition, called LBC-condition, is that in the weak/strong definition, the time complexity must be some polynomial of n and log m. In this paper, it is shown that under these conditions, there still exist learners which execute exhaustive search. For instance, there exists a learner which learns CFL represented by CFG in the weak definition using membership queries and equivalence queries under the proper termination condition, and there also exists a learner that learns CFL represented by CFG in the strong definition using subset queries and superset queries under LBC-condition. These results suggest that the weak definition is not an adequate learning model even if the proper termination condition is assumed. Also, the model becomes inadequate in the strong definition if some combination of queries, such as subset queries and superset queries, is used instead of equivalence queries. Many classes of languages become learnable by our extracting long counterexample" technique. However, it is still open whether or not CFL represented by CFG is learnable in the strong definition from membership queries and equivalence queries, although the answer is known to be negative if at least one of (1) quadratic residues modulo a composite, (2) inverting RSA encryption, or (3) factoring Blum integers, is intractable.

Almost stoichiometric YBa2Cu3O7-δ(110) or (103) and SrTiOx(110) films, and multilayer films consisting of them have successfully been grown epitaxially on hot SrTiO3 substrates by 90off-axis rf magnetron sputtering with facing targets. Their whole composition, compositional distribution in depth, crystallinity and surface morphology were examined by inductively coupled plasma spectroscopy, Auger electron spectroscopy, reflection high-energy electron diffraction, and scanning tunneling microscopy or atomic force microscope, respectively. When any YBa2Cu3O7-δ film was exposed to air after deposition, a Ba-rich layer was formed in a near surface region of the film. However, such a compositional distribution in depth of the film was improved by in situ deposition of a SrTiOx film on it. Moreover, the surface roughness of the YBa2Cu3O7-δ film was improved by predeposition of a SrTiOx film under it. On the basis of these results, both YBa2Cu3O7-δ/SrTiOx/YBa2Cu3O7-δ/SrTiO3(sub.) and YBa2Cu3O7-δ/SrTiOx/YBa2Cu3O7-δ/SrTiOx/SrTiO3(sub.) multilayer films with average surface roughness of 3 nm were grown reproducibly, which had uniform compositional distribution throughout the depth of the film except a near surface region of the top YBa2Cu3O7-δ layer. A new 222 structure described by Sr8Ti8O20 (Sr2Ti2O5) with a long range ordered arrangement of oxygen vacancies was formed in the SrTiOx films deposited epitaxially on YBa2Cu3O7-δ films.

This paper describes a pipelining universal system of discrete time cellular neural networks (DTCNNs). The new relaxation-based algorithm which is called a Pipelining Gauss Seidel (PGS) method is used to solve the CNN state equations in pipelining. In the systolic system of N processor elements {PEi}, each PEi performs the convolusional computation (CC) of all cells and the preceding PEi-1 performs the CC of all cells taking precedence over it by the precedence interval number p. The expected maximum number of PE's for the speeding up is given by n/p where n means the number of cells. For its application, the encoding and decoding process of moving images is simulated.

Overlapping multi-pulse pulse position modulation (OMPPM) is a modulation scheme having higher capacity and cutoff rate than other conventional modulation schemes when both off-duration between pulses shorter than a laser pulsewidth and resolution better than a laser pulsewidth are realized [1],[2]. In Refs. [1],[2] erasure events of a few chips that can be decoded correctly is defined as an erasure event. This results in lower bounds on the performance of OMPPM in optical-direct-detection channel in quantum limited case. This paper analyzes more exact performance of OMPPM in optical direct-detection channel in quantum limited case when both off-duration between pulses shorter than a laser pulsewidth and resolution better than a laser pulsewidth are realized. First we derive the error probability of OMPPM with considering what chips are detected or erased. Then we derive the capacity and the cutoff rate of OMPPM using the error probability. It is shown that OMPPM outperforms on-off keying (OOK), pulse position modulation (PPM), multi-pulse PPM (MPPM), and overlapping PPM (OPPM) in terms of both capacity and cutoff rate for the same pulsewidth and the same duty cycle. Moreover, it is shown that OMPPM with fewer slots and more pulses per block has better cutoff rate performance when the average received power per slot is somewhat large.

This paper describes highly parallel analog image coding and decoding by cellular neural networks (CNNs). The communication system in which the coder (C-) and decoder (D-) CNNs are embedded consists of a differential transmitter with an internal receiver model in the feedback loop. The C-CNN encodes the image through two cascaded techniques: structural compression and halftoning. The D-CNN decodes the received data through a reconstruction process, which includes a dynamic current distribution, so that the original input to the C-CNN can be recognized. The halftoning serves as a dynamic quantization to convert each pixel to a binary value depending on the neighboring values. We approach halftoning by the minimization of error energy between the original gray image and reconstructed halftone image, and the structural compression from the viewpoints of topological and regularization theories. All dynamics are described by CNN state equations. Both the proposed coding and decoding algorithms use only local image information in a space inveriant manner, therefore errors are distributed evenly and will not introduce the blocking effects found in DCT-based coding methods. In the future, the use of parallel inputs from on-chip photodetectors would allow direct dynamic quantization and compression of image sequences without the use of multiple bit analog-to-digital converters. To validate our theory, a simulation has been performed by using the relaxation method on an 150 frame image sequence. Each input image was 256256 pixels whth 8 bits per pixel. The simulated fixed compression rate, not including the Huffman coding, was about 1/16 with a PSNR of 31[dB]35[dB].

In this paper we propose a new algorithm for recognizing 3-D objects from 2-D images. The algorithm takes the multiple view approach in which each 3-D object is modeled by a collection of 2-D projections from various viewing angles where each 2-D projection is called an object model. To select the candidates for the object model that has the best match with the input image, the proposed algorithm computes the surface matching score between the input image and each object model by using Hopfield nets. In addition, the algorithm gives the final matching error between the input image and each candidate model by the error of the pose-transform matrix proposed by Hong et al. and selects an object model with the smallest matching error as the best matched model. The proposed algorithm can be viewed as a combination of the algorithm of Lin et al. and the algorithm of Hong et al. However, the proposed algorithm is not a simple combination of these algorithms. While the algorithm of Lin et al. computes the surface matching score and the vertex matching score berween the input image and each object model to select the candidates for the best matched model, the proposed algorithm computes only the surface matching score. In addition, to enhance the accuracy of the surface matching score, the proposed algorithm uses two Hopfield nets. The first Hopfield net, which is the same as that used in the algorithm of Lin et al., performs a coarse matching between surfaces of an input image and surfaces of an object model. The second Hopfield net, which is the one newly proposed in this paper, establishes the surface correspondences using the compatibility measures between adjacent surface-pairs of the input image and the object model. the results of the experiments showed that the surface matching score obtained by the Hopfield net proposed in this paper is much more useful for the selectoin of the candidates for the best matched model than both the sruface matching score obtained by the first Hopfield net of Lin et al. and the vertex matching score obtained by the second Hopfield net of Lin et al. and, as the result, the object recognition algorithm of this paper can perform much more reliable object recognition than that obtained by simply combining the algorithm of Lin et al. and the algorithm of Hong et al.

We define the complexity and the distortion-complexity of an individual finite length string from a finite set. Assuming that the string is produced by a stationary ergodic source, we prove that the distortion-complexity per source letter and its expectation approximate arbitrarily close the rate-distortion function of this source as the length of the string grows. Furthermore, we apply this property to construct a universal data compression scheme with distortion.

This paper propeses a new motion compensation (MC) technique which reduces blurring called drift in moving pictures down-scaled with layered coding system. Encoder of the system compresses large amounts of digital video data in the same way of MPEG (Moving Picture Experts Group) algorithm. Decoder, on the other hand, expands a part of the compressed data and reconstructs down scaled pictures. The purpose of this paper is to reduce blurring which is observed in the reconstructed pictures. In this paper, cause of the blurring is analyzed and the method is introduced as a solution to the problem. The new method is implemented by a little modification of motion compensation (MC) of the decoder, namely increasing the number of tap of interpolation fillters of the MC. Compressing moving pictures, its effectiveness is also confirmed by means of not only subjective test but also signal to noise ratio.

A feedback pattern recognition method using an inverse recall neural network model is proposed. The feedback method can adjust processing parameter values adaptively to individual patterns so as to produce reliable recognition results. In order to apply an adaptive control technique to such pattern recognition processings, the evaluation value for recognition uncertainty is determined to be a function with regard to an input pattern and processing parameters. In its feedback phase, the input pattern is fixed and processing parameters are adjusted to decrease the recognition uncertainty. The proposed neural network model implements two functions in this feedback recognition method. One is a discrimination as a kind of multi-layer feedforward model. The other is to generate an input modification so as to decrease the recognition uncertainty. The modification values indicate parts which are important for more certain recognition but are missed in the original input to the nerwork. The proposed feedback method can adjust prcessing parameter values in order to detect the important parts shown by the inverse recall network model. As explained in this paper, feature extraction parameter values are adaptively adjusted by this feedback method. After the inverse recall model and the feedback function are implemented, features are extracted again by using the modified feature extraction parameter values. The feature is classified by the feedforward function of the network model. The feedforward and feedback processings are repeated until a certain recognition result is obtained. This method was examined for hadwritten alpha-numerics with rotation distortion. The feedback method was found to decrease the rejection ratio at the same substitution error ratio with high efficiency.

This paper reviews two topics of nonlinear system analysis done in Japan. The first half of this paper concerns with nonlinear system analysis through the nondeterministic operator theory. The nondeterministic operator is a set-valued or fuzzy set valued operator by K. Horiuchi. From 1975 Horiuchi has developed fixed point theorems for nondeterministic operators. Using such fixed point theorems, he developed a unique theory for nonlinear system analysis. Horiuchi's theory provides a fundamental view point for analysis of fluctuations in nonlinear systems. In this paper, it is pointed out that Horiuchi's theory can be viewed as an extension of the interval analysis. Next, Urabe's theory for nonlinear boundary value problems is discussed. From 1965 Urabe has developed a method of computer assisted existence proof for solutions of nonlinear boundary value problems. Urabe has presented a convergence theorem for a certain simplified Newton method. Urabe's theorem is essentially based on Banach's contraction mapping theorem. In this paper, reformulation of Urabe's theory using the interval analysis is presented. It is shown that sharp error estimation can be obtained by this reformulation. Both works discussed in this paper have been done independently with the interval analysis. This paper points out that they have deep relationship with the interval analysis. Moreover, it is also pointed out that these two works suggest future directions of the interval analysis.

We study robot navigation in unknown environment with rectangular obstacles aligned with the x and y axes. We propose a strategy called the modified-bian heuristic, and analyze its efficiency. Let n be the distance between the start point and the target of robot navigation, and let k be the maximum side length among the obstacles in a scene. We show that if k=(o(n) and if the summation of the widths of the obstacles on the line crossing the target and along the y axis is o(n), then ratio of the total distance walked by the robot to the shortest path length between the start point and the target is at most arbitrarily close to 1+k/2, as n grows. For the same restrictions as above on the sizes of the obstacles, the ratio is also at most arbitrarily close to 1+3/4n, as n grows, where is the summation of lengths of the obstacles in y axis direction.

An Ordered Binary Decision Diagram (BDD) is a graph representation of a Boolean function. According to its good properties, BDD's are widely used in various applications. In this paper, we investigate the computational complexity of basic operations on BDD's. We consider two important operations: reduction of a BDD and binary Boolean operations based on BDD's. This paper shows that both the reduction of a BDD and the binary Boolean operations based on BDD's are NC1-reducible to REACHABILITY. That is, both of the problems belong to NC2. In order to extend the results to the BDD's with output inverters, we also considered the transformations between BDD's and BDD's with output inverters. We show that both of the transformations are also NC1-reducible to REACHBILITY.

Finite state translation systems (fsts') are a widely studied computational model in the area of tree automata theory. In this paper, the string generating capacities of fsts' and their subclasses are studied. First, it is shown that the class of string languages generated by deterministic fsts' equals to that of parallel multiple context-free grammars, which are an extension of context-free grammars. As a corollary, it can be concluded that the recognition problem for a deterministic fsts is solvable in O(ne1)-time, where n is the length of an input word and e is a constant called the degree of the deterministic fsts'. In contrast to the latter fact, it is also shown that nondeterministic monadic fsts' with state-bound 2 can generate an NP-complete language.

We present outside-in conditional narrowing for orthogonal conditional term rewriting systems, and show the completeness of leftmost-outside-in conditional narrowing with respect to normalizable solutions. We consider orthogonal conditional term rewriting systems whose conditions consist of strict equality only. Completeness results are obtained for systems both with and without extra variables. The result bears practical significance since orthogonal conditional term rewriting systems can be viewed as a computation model for functional-logic programming languages and leftmost-outside-in conditional narrowing is the computing mechanism for the model.

There have been several interesting investigations on the space functions constructed by one-dimensional or two-dimensional Turing machines. On the other hand, as far as we know, there is no investigation about the space functions constructed by three-dimensional Turing machines. In this paper, we investigate about space constructibility by three-dimensional deterministic Turing machines with cubic inputs, and show that the functions log*n and log(k)n, k1, are fully space constructible by these machines.

We developed an automatic data processing algorithm for a ground-probing radar which is essential in analyzing a large amount of data by a non-expert. Its aim is to obtain an optimum result that the conventional technique can give, without the assistance of an experienced operator. The algorithm is general except that it postulates the existence of at least one isolated target in the radar image. The raw images of underground objects are compressed in the vertical and the horizontal directions by using a pulse-compression filter and the aperture synthesis technique, respectively. The test function needed to configure the compression filter is automatically selected from the given image. The sensitivity of the compression filter is adjusted to minimize the magnitude of spurious responses. The propagation velocity needed to perform the aperture synthesis is determined by fitting a hyperbola to the selected echo trace. We verified the algorithm by applying it to the data obtained at two test sites with different magnitude of clutter echoes.

Let S(n) be a space constructible function such that S(n) log n. In this paper, we show that AuxSpTu (S(n),T(n)) NSPACE (S(n)log T(n)), where AuxSpTu (S(n),T(n)) is the class of languages accepted by nondeterministic auxiliary pushdown automata operating simultaneously in O(S(n)) space and O(T(n)) turns of the auxiliary tape head.

This paper describes a nondestructive measurement method for complex permittivity of dielectric material at pseudo microwave frequencies. The resonator used in this study has a cylindrical cavity filled with a sapphire material of a well known complex permittivity. The resonator is divided into two parts at the center. A dielectric substrate specimen is clamped with these halves. Relative permittivity εand loss tangent tan δ of the specimen are obtained at 3 GHz using the TE011 resonance mode. The accuracy of the present method is evaluated through the comparison of the measured values by the new method with those at around 10 GHz by the conventional empty cavity resonator method. The errors of measurements are smaller than 1% and 1105 for εand tan δ, respectively.

A noninvasive method for measuring complex permittivity of biological tissues is proposed. The noninvasive method is based on an inverse scattering technique which employs an iterative procedure. The iterative procedure consists of solving an electric field integral equation using the method of moments and minimizing the square difference between calculated and measured scattered fields. Implementation of the noninvasive method requires the knowledge of the target shape, the incident and measured scattered fields. Based on the noninvasive method, a measuring system of complex permittivity is developed and its reliability is verified.