Web page visibility gained importance due to the rapid dissemination of the World Wide Web. Since combinations of foreground and background colors are crucial in providing good visibility and an investigation of all possible color combinations is extremely difficult to conduct, a model or empirical formula for estimating color combination visibility is needed. In this study, a visibility rating method of color combinations is proposed. The visibilities of several web-safe color combinations were examined using a psychological method, a paired comparison. Twenty students with normal color sensations ranging from ages 19 to 26 were recruited. They were instructed to look at two different colored rings simultaneously on the same background and identify which one enabled them to see better. When examining the relationship between the psychological rankings of color combinations and visual sensations, the visibility of each color combination was first scored by Thurstone's paired comparisons technique. The visual sensations were then deduced by applying Weber-Fechner's law to the luminance of the stimuli. We found that color combination ranking has a positive monotonic relationship with sensation magnitude. This indicates that the visibilities of color combinations can be rated by applying Weber-Fechner's law to the luminance pairs of foreground and background colors.

抄録全体を表示

In this paper, we present an application of soft computing into an expert system RETS: Rheumatic Evaluation and Treatment System by Oriental Medicine (OM). Inputs are severities of observed symptoms on patients and outputs are a diagnosis of rheumatic states, its explanation and herbal prescriptions. First, after using fuzzy inferences, RETS diagnoses the most suitable rheumatic state in which the patient appears to be infected and shows explanations. Next, based on prescriptions for rheumatism learned by neural networks from skilled OM physicians and OM text books, RETS gives an oriental prescription written in suitable herbs with reasonable amounts. Finally, we describe evaluations and restrictions of RETS, and then describe our future work.

抄録全体を表示

Dental radiographs are of immense help in the identification, and evaluation of oral pathologies. One of the common oral pathology is a dental cyst. Objective quantification of severity and early stage detection highly benefits the diagnosis and treatment of the dental cysts. In this paper, we propose a novel neural network based automated system to identify and quantify the severity of the cysts using dental radiograph images. An automated diagnosis of dental cyst in radiography images based on segmentation algorithm and Artificial Neural Networks (ANN) is presented. By using template-matching approach templates pertaining to various cysts are slide over the input image to obtain the Normalized Cross Correlation (NCC) and Extended Normalized Cross Correlation (ENCC) images. An ANN is trained using ENCC values to locate the suspicious region. The diagnosis of the cysts is brought out by the extraction of connected components in the original image. The results obtained provide details about severity of the cysts and thereby increases the diagnostic ease of dental surgeon.

抄録全体を表示

In this study, we propose multilayer neural networks with adjustable intermediate elements and show their higher learning capability by computer simulation. A well-known backpropagation learning algorithm has a problem because of local minima. In the previous work, Yokoi et al. have proposed intermediate elements, which intervene between the two layers and function to perform feature detection and categorization, in order to avoid increasing of local minima and to increase the learning capability. They have demonstrated that the learning capability is increased due to intermediate elements. However, evaluation of intermediate elements under the optimum learning condition and adjusting of intermediate elements have not yet been discussed. The optimum learning condition is needed to adequately evaluate the best performance of neural networks. Adjusting of intermediate elements is expected to more increase the learning capability though the design of networks becomes complex. In this paper, we presented a Design of Experiments based optimization method to obtain the optimum learning condition such as the number of elements in hidden layers, a learning rate and momentum, and showed that the learning capability in the case of using adjustable intermediate elements is much higher than constant intermediate elements.

抄録全体を表示

We proposed the method of complex diffusion equation. Then, we started with a general stochastic deferential equation, and got to a general complex equation which is similar to wave equation of quantum mechanics. If a change of the floating diffusion coefficient is so small, the complex equation is almost equivalent to Schrodinger Equation. We showed that it is safety to adopt a same complex diffusion equation if there are the common fundamental phenomena as Markov process or stochastic dynamics. The concept of the scaling law is related with a renormalization group. An expectation value and dispersion were defined by the wave function of the complex diffusion equation. The method of the complex diffusion equation (MCDE) is a generalization of functional data analysis. We think that minimum uncertainties have essentially nothing to do with the fuzziness of statistical method. At the next step, we have shown a new application of the system analysis, where probability density function obeys the wave equation or Fokker-Planck equation. We consider the fuzziness of the data by introducing fuzzy derivative, we can estimate the width of the predicted data by the fuzzy integration. This method does not assume the shape of the population of sample data. In this method, every knowledge of the system is accumulated in the real and imaginary potentials, so if we accumulate the knowledge, we can read the trend of the past data or foretell the future events

抄録全体を表示

Generally, we could efficiently analyze the inexact information and investigate the fuzzy relation by applying the fuzzy graph theory. We would extend the fuzzy graph theory, and propose a fuzzy node fuzzy graph. Since a fuzzy node fuzzy graph is complicated to analyze, we would transform it to a simple fuzzy graph by using T-norm family. In this paper, we would discuss about four subjects, (1) fuzzy node fuzzy graph, (2) new T-norm "quasi logical product", (3) decision analysis of the optimal fuzzy graph G_<λ0> in the fuzzy graph sequence {G_λ} and (4) Us application to sociometry analysis. By using the fuzzy node fuzzy graph theory and the new T-norm, we could clarify the relational structure of fuzzy information. According to the decision method in section 2, we could find the optimal fuzzy graph G_<λ0> in the fuzzy graph sequence {G_λ}, and clarify the structural feature of the fuzzy node fuzzy graph. Moreover, we would illustrate its practical effectiveness with the case study concerning sociometry analysis.

抄録全体を表示

Classical molecular dynamics simulations are used to study a possibility of storing guest molecules in clathrate hydrate. In our calculation, the clathrate hydrate units are taken to be rigid and the intermolecular interactions are modeled as a sum of van der Waals interactions with parameters from AMBER force. The initial structure of clathrate is constructed on the symmetry of Fd3m, and static properties are investigated. The time dependent motions of with H_2 guest gas molecules is determined by the calculation of molecular dynamics. We found the movement of H_2 gas from one to another large cage are often occurred through the hole of the large cages, while seldom in small cages, and the frequency of movement goes up according to the temperature. It is concluded that the H_2 gas are stored under low temperature and high pressure.

抄録全体を表示