|
1. |
Relational Matching Using a Neural Network |
|
Systems and Computers in Japan,
Volume 21,
Issue 6,
1990,
Page 1-9
Noboru Katta,
Hideya Takahashi,
Kenji Matsushita,
Eiji Shimizu,
Preview
|
PDF (549KB)
|
|
摘要:
AbstractThis paper proposes an algorithm for relational matching, which is important in pattern recognition, by using the Hopfield model as one of the neural network models. First, the objective function for relational matching is derived and the computer simulation of the neural networking for relational matching according to this objective function is conducted by use of a simple model. Next, the local minima in connection with the Hopfield model is avoided by using triangular noise and its effectiveness is verified in the network for relational matching. This method differs from the conventional simulated annealing method and, by adding noise to the network, the network will reach only one stationary state for which results will be derived by this method.
ISSN:0882-1666
DOI:10.1002/scj.4690210601
出版商:Wiley Subscription Services, Inc., A Wiley Company
年代:1990
数据来源: WILEY
|
2. |
Estimation for Spread Forms of Spindle Waves in Multichannel EEG |
|
Systems and Computers in Japan,
Volume 21,
Issue 6,
1990,
Page 10-20
Yoshikazu Ueda,
Naohiro Ishii,
Preview
|
PDF (814KB)
|
|
摘要:
AbstractThe spindle wave is one of the most important EEG signals in exploring the mechanism of sleep. There have been few reports, however, that examined the spatial properties of the spindle wave. One reason for this may be that the spindle wave is difficult to be handled as a stationary time series since it is a short‐term signal. Another reason is that a multidimensional signal processing technique has not been established for the multichannel signal, and an analysis with a tremendous amount of data has not been tried.This paper aims at the analysis of the empirical knowledge of doctors by computer processing, and proposes an algorithm which estimates the spread form of the spindle wave in the multichannel EEG. The method is based on the frequency‐wavenumber spectrum, and has features in that the two phasic dominants and the direction of spread are estimated from a smaller number of measurements, by utilizing the phase relations on the scope.The usefulness of the proposed algorithm was verified by a computer simulation. The proposed estimation method was applied to the actual spindle waves, and the following result was obtained. The phasic dominant of the spindle exists around the central field for the low‐frequency components and around the parietal field for high‐frequency components. This is a quantitative result, which agrees with the traditional qualitative findings. A new observation was also made where the spindles spread mostly from the parietal field to the fronta
ISSN:0882-1666
DOI:10.1002/scj.4690210602
出版商:Wiley Subscription Services, Inc., A Wiley Company
年代:1990
数据来源: WILEY
|
3. |
Design of an RSA Encryption Processor Based on Signed‐Digit Multivalued Arithmetic Circuits |
|
Systems and Computers in Japan,
Volume 21,
Issue 6,
1990,
Page 21-31
Michitaka Kameyama,
Shugang Wei,
Tatsuo Higuchi,
Preview
|
PDF (717KB)
|
|
摘要:
AbstractThis paper proposes an LSI‐oriented multiple‐valued hardware algorithm based on the signed‐digit number (SD number) for highspeed RSA public key encryption processor. A very long word‐length arithmetic is required in RSA encryption processing. The proposed algorithm realizes a high‐speed processing by iterative additions based on the radix‐A SD number system. Microprogram control is adopted in the encryption processor. To verify the basic operation of the proposed processor architecture, a model processor for 16‐bit encryption is implemented using binary TTL logic IC's. A comprehensive evaluation of the 2 μm CMOS encryption LSI is made by a simulation using the electronic circuit analysis program SPICE2. The encryption rate for the message of 512‐bit word length is 60 kbit/s, which is a factor of eight times faster than that of the corresponding binary LSI. The developed processor contains about 100 k transistors and it can be implemented on a si
ISSN:0882-1666
DOI:10.1002/scj.4690210603
出版商:Wiley Subscription Services, Inc., A Wiley Company
年代:1990
数据来源: WILEY
|
4. |
Improvement of HDTV‐PAL Standard Converter Using Motion‐Compensation Technique and Detection Probability of Motion Vectors |
|
Systems and Computers in Japan,
Volume 21,
Issue 6,
1990,
Page 32-42
Yutaka Tanaka,
Taiji Nishizawa,
Toshiro Ohmura,
Preview
|
PDF (753KB)
|
|
摘要:
AbstractHDTV signal can be converted into the present standard TV signal with a high spatial resolution. However, some problems have been pointed out in the standard conversion with different field frequencies, which are the degradation of resolution and jitter for moving pictures. To avoid such problems, the motion compensation frame conversion method was developed, which realized a practically satisfactory standard conversion. In the development of a practical system for the standard conversion from HDTV to PAL, the quality of the converted picture is improved by improving the detection of motion vector and the processing of motion‐compensated picture. The detection probability for the motion vector is investigated, and the usefulness of the motion compensation based on the detection of the motion vector used in the proposed system is verifie
ISSN:0882-1666
DOI:10.1002/scj.4690210604
出版商:Wiley Subscription Services, Inc., A Wiley Company
年代:1990
数据来源: WILEY
|
5. |
Filtering for Dynamic Systems Under Unknown State transition Matrices |
|
Systems and Computers in Japan,
Volume 21,
Issue 6,
1990,
Page 43-50
Katsumi Sakata,
Kiichiro Izumida,
Preview
|
PDF (527KB)
|
|
摘要:
AbstractA filter is proposed which generates the state estimate independently of the physical condition which determines the dynamic characteristics of a system. This filter consists in digital first‐order filters and a controller. The predicted state is corrected repeatedly until the weighted mean of the predicted residual square becomes less than the residual covariance. The correction procedure is that the predicted residual is input to digital first‐order filters, the outputs of these filters are multiplied by constants and the result is added to the predicted state as a correc
ISSN:0882-1666
DOI:10.1002/scj.4690210605
出版商:Wiley Subscription Services, Inc., A Wiley Company
年代:1990
数据来源: WILEY
|
6. |
A Reconstructed Image Analysis and a Practical Algorithm of the Direct 2‐D Fourier Transform |
|
Systems and Computers in Japan,
Volume 21,
Issue 6,
1990,
Page 51-63
Noboru Niki,
Preview
|
PDF (978KB)
|
|
摘要:
AbstractSignificant research and development have been made on various kinds of CT's such as the X‐ray, the MRI and the positron, which are applied widely in practice. The CT device is divided into the measurement system and the image reconstruction system. Efforts are still being made at present to improve the resolution and speed of those systems. This paper discusses the 2‐D Fourier transform method, which is one of the image reconstruction methods that have been investigated and applied in the image reconstruction in various types of CT's.Traditionally, the 2‐D Fourier transform method has been considered as suffering from low image quality, although the image can be reconstructed with a high speed. The reason for image degradation is the lack of accuracy in the interpolation in the 2‐D Fourier transform domain, which can be remedied by improving the accuracy.A problem then is that the coordinate transformation in the Fourier domain depends on the individual CT. From such a viewpoint, this paper analyzes the relation between the coordinate transformation and the reconstructed image to indicate the reason for the image degradation.A systematic analysis is made of the reconstructed image when the linear interpolation is applied to each coordinate transformation, to indicate the reason for the image degradation and to describe the artifact appearing in the reconstructed image. Based on the result of this analysis, an algorithm is constructed which can eliminate the artifact from the reconstructed image. The algorithm is composed of the normalization algorithm for the interpolation characteristics, the interpolation by the fast‐Fourier transform and the inverse 2‐D Fourier transform enlarging the reconstructed image region. The effectiveness of the method is demonstrated by applying the method to the transformation from the polar to the Cartesian
ISSN:0882-1666
DOI:10.1002/scj.4690210606
出版商:Wiley Subscription Services, Inc., A Wiley Company
年代:1990
数据来源: WILEY
|
7. |
Knowledge Level Learning in a Problem Solver |
|
Systems and Computers in Japan,
Volume 21,
Issue 6,
1990,
Page 64-70
Seiichiro Sakurai,
Masamichi Shimura,
Preview
|
PDF (473KB)
|
|
摘要:
AbstractThis paper describes a new knowledge level learning method for a problem solver. Because it is not practical to initialize a problem solver with a very large knowledge base to cope with a large variety of problems, the machine must have the capability of acquiring knowledge through solving problem in an incremental fashion. New knowledge is acquired during dialog with an instructor in problem solving operations. The accumulation of experience in problem solving enables the generalization of rules and knowledge for multiple problems. Because each new rule is formulated after solving some problem and then verified by the instructor the knowledge base will be more robust. In addition, errors in the knowledge base can be corrected by the instructor during the dialog.
ISSN:0882-1666
DOI:10.1002/scj.4690210607
出版商:Wiley Subscription Services, Inc., A Wiley Company
年代:1990
数据来源: WILEY
|
8. |
Directional Feature Field and Parallel Operations for Binary Images: Twisting Operation for Arbitrary Directional Propagation by 8‐Neighbors |
|
Systems and Computers in Japan,
Volume 21,
Issue 6,
1990,
Page 71-80
Hiromitsu Yamada,
Kazuhiko Yamamoto,
Preview
|
PDF (742KB)
|
|
摘要:
AbstractThis paper points out the importance of directional propagation with an expression of directional elements at each pixel of a binary image, as an extension of local parallel operations. This paper also proposes “a twisting operation” for an 8‐neighbor operation with an arbitrary direction and an arbitrary directional accuracy. By using this operation and the concept of vibrating operations, such as erosion‐dilatation or dilatation‐erosion, directional local parallel operations are defined to extract linear features. The effectiveness of the ideas is confirmed exper
ISSN:0882-1666
DOI:10.1002/scj.4690210608
出版商:Wiley Subscription Services, Inc., A Wiley Company
年代:1990
数据来源: WILEY
|
9. |
Establishing Correspondence and Obtaining 3‐D Information in Dynamic Images |
|
Systems and Computers in Japan,
Volume 21,
Issue 6,
1990,
Page 81-94
Kaoru Suzuki,
Masahiko Yachida,
Preview
|
PDF (848KB)
|
|
摘要:
AbstractTo establish correspondences between dynamic images, methods based on their local similarity such as the correlation method have been used widely. If corresponding points are searched over the whole image, however, the computation time and risk of occurrence of false correspondences increase. Therefore, it is important to restrict the search area.This paper proposes a reliable and efficient method of establishing correspondences among the sequences of dynamic images by using geometrical constraint conditions derived from the positional relationship of viewpoints before and after a movement. Also proposed is another method by which the correspondences between images formed by the binocular view of a mobile robot and the correspondences between images formed before and after a movement can be carried out simultaneously. The experimental results show that the correspondences of dynamic stereo images and the range information are obtained successfully.
ISSN:0882-1666
DOI:10.1002/scj.4690210609
出版商:Wiley Subscription Services, Inc., A Wiley Company
年代:1990
数据来源: WILEY
|
10. |
Automated Design of Circuits from Recursion Equations Using Theorem‐Proving Technique |
|
Systems and Computers in Japan,
Volume 21,
Issue 6,
1990,
Page 95-107
Masateru Harao,
Kouji Iwanuma,
Preview
|
PDF (908KB)
|
|
摘要:
AbstractThis paper aims at establishing the automated design for a circuit by the theorem‐proving technique, by formulating the circuit design as a transformation from the specification description (recursion equation) to the circuit model FN (functional network based on dataflow operation).First, rules for circuit construction and transformation rules are formulated, which are used as the inference knowledge in the theorem‐proving system. Then the circuit realization language is designed based on the modal logic, where the syntax corresponds to the FN structure, and the computation corresponds to the circuit operation. It is shown that the processing system for such a language can be realized by a technique similar to the derivation principle for the first‐order predicate logic.The theorem‐proving system for automated circuit design can be realized as a program on the circuit realization language, where the element composing the circuit is used as the declarative knowledge, and the transformation rule is used as the inference knowledge.When the design specification (recursion equation) is given as the goal clause, the automated circuit design system decides on the possibility of design, and the circuit structure information is extracted from the proof process. The behavior of the circuit can be simulated also in the
ISSN:0882-1666
DOI:10.1002/scj.4690210610
出版商:Wiley Subscription Services, Inc., A Wiley Company
年代:1990
数据来源: WILEY
|
|