Esthetic improvement of the faceis one of the goals of orthognathic surgery. In evaluation of the esthetic outcome, profile is subjectively evaluated through visual cognition or cephalometric soft tissue profile analysis. Although cephalometric variables provide a numerical foundation for profile assesment, overall evaluation depends on a clinician's subjective jedgement. Therefore, a reproducible index is desirable for clinical application. We developed an esthetic evaluation system of the profile using fuzzy rule-based inference, which produced consistent correlation between the subjective ratings and the selected cephalometric variables.

View full abstract

This paper studies a knowledge-based system implementation method to realize the preliminary diagnosis of nasal allergy. Constructing the rule base and the case base, respectively based on the knowledge of doctor and the questionnaire of patient at first, and putting the information of the patient for diagnosis into the two bases and scanning the bases, the system can give the suitable diagnosis. The rule-base reasoning and learning (RBRL) algorithm are performed by a neural network at first, and then the case base reasoning (CBR) algorithm is also performed by using fuzzy method. On the other hand, since algorithm of the neural networks can be convertible to that of fuzzy method, it is reasonable to combine the merits of both paradigms, that is, to combine the learning ability of neural network and the easy implementation of fuzzy logic.

View full abstract

The Fuzzy set cloning (FSC) theory was developed to identify new functional protein(s) easily and rapidly. Characteristically, in FSC the functional target protein of interest is reported as a cDNA sequence. Using FSC, most of the proteins, if they configure mass to function, are "harvested" in a "net" using fuzzy set antibody (FSAb). The cDNAs are then subjected to a specific selection procedure called defuzzification (DeFuzzy). Here, the most useful fast procedure with explicit object function should be selected for DeFuzzy. The defuzzified clones are sequenced and BLAST / IDEAS program is used to determine whether they represent a new or useful functional protein by comparing the sequences to those in the EMBL and GenBank database. We focused on inducer proteins of rat liver fibrosis after injecting the animals with carbon tetrachloride (CCl_4). Using antibody against basement membrane as FSAb, 32 positive λgt11 cDNA clones were obtained. For DeFuzzy, gene expression levels were analyzed on timed Northern blots. Sequencing of a clone coding mRNA expressed in the early phase after CCl_4 treatment identified human profilin. The profilin peptide antibody stained fibrogenic areas in cirrhotic livers from rats.

View full abstract

Real three-dimensional features of human cerebellar cortex were observed using scanning electron microscopy (SEM). A specimen preparation method for SEM in the present study was a t-butyl alcohol freeze-drying device at 15℃ and 160mmHg. Relatively large specimens (about 10 x 15 mm x 1 mm) of formalin-fixed human, cut with a razor blade, were rinsed in water, dehydrated in a series of ethanols, immersed in a series of t-butyl alcohol, and then placed in the new freeze-drying device. The specimens were freeze-dried without acid or alkali digestion, mounted on stubs with a diameter of 32 mm, and sputter-coated with gold. This new preparation method allowed a higher magnification of surfaces of cells and fibers of the human cerebellar cortex compared to the critical point drying method. This was valied for Purkinje cell bodies with axons, dendrites with climbing fibers and climbing fiber glomeruli, stellate neuron cell processes connected to the Purkinje cell dendrites. Lugaro cell, Golgi II cell, basket cells with axons, mossy fiber glomerulus with granule cell dendrites, ell dendrites, satellite Bergmann glial cells with processes and many microtubule-like fibrous structures on the inside of Purkinje cell dendrites are shown.

View full abstract

Fractal theory provides an effective method of quantitatively describing complicated shapes and patterns in nature. We applied fractal theory for analyzing computed tomography (CT) images. Raw data of upper abdominal CT scanning performed on normal, cirrhotic and fatty livers were transferred to a personal computer, the images were reproduced on the computer, and regions of interest (ROIs) were set on the right lobe of the liver. The raw CT numbers within the ROIs were then calculated to obtain fractal dimensions using a scale conversion method. As a result, CT of the liver, which has a relatively uniform texture pattern, showed fractality. Fractal dimensions of the fatty liver were significantly lower than that of the normal liver. Fractal dimension analysis may provide a quantitative measure of parenchymal patterns of the liver on CT.

View full abstract

The two theories concerning mixed emotions, adopted in this research is briefly explained. Also discussed are the image codes linked to the mixed emotional words. Next, we explain the theory for Subjective Observation Model (SOM)[2], which have high applicability to several kinds of fields especially on the emotional processing [1][2]. This paper proposes a basic concept of human cognitive model based on the theory for SOM and shows an example of application to emotional processing. We refer to the mapping function which forms our cognitive model, and show some emotional maps which were drawn by computer simulation, which seem like "Screen of Consciousness" in our mind.

View full abstract

This paper presents a comparative study of expert system, fuzzy logic system and neural networks in the detection of coronary artery disease. It is evident from this study that expert system can be useful in training and providing expert assistance to the user. Fuzzy logic system simplifies the knowledge base in the system. Neural networks are faster and have generalisation capability.

View full abstract

The line drawing is generated from input image by using neural network as feature detection. The direction of edges of the object in the image is detected as a local feature. The primitive pattern is drawn on output image according to the detected local features. The various output image is generated by changing the some parameters. This system consists of feature detection part and drawing part. We describe of a method of making the training set by which the weights of neural network is learned so as to detect the appropriate features. The drawings by using this system and the detected local features are shown.

View full abstract

The sociometry is one of the measurement and evaluation methods of social structure which we could effectively analyze by applying fuzzy graph. According to the data obtained from the simple questionnaires, we could measure the fuzzy relations among the members of a group and observe its human structure by applying Shapley value. In this paper, we would not only explain an analysis method of fuzzy sociograms, but also illustrate its practical effectiveness with case study from psychological and biomedical points of view.

View full abstract

It is important to prevent the surgical site infection after operation. Surgeons diagnose it be using many kinds of patient's information such as fever, counts of white blood cells, the complains of pain and use the antibiotics to treat it. But it is difficult to evaluate the effect of the medicine, because these patient's information has the fuzziness and the diagnostic logic of the surgeon is subjective. The purpose of this paper is to propose a fuzzy logic for detecting the surgical site infection, employing the fuzzy inference based on the data of differential counts of white blood cells. By the application of this logic to the clinical cases, we had the same results as the ones by the discriminate analysis with many variables.

View full abstract

Assessment of ambulatory blood pressure monitoring (ABPM) was approached by fuzzy measurements. Four measurement items obtained from ABPM were quantified with fuzzy models. The items were 1) mean systolic blood pressure, 2) hyperbaric area (HBA), 3) maximal systolic blood pressure, 4) minimal systolic pressure (fig.1-4). Weights among items in comprehensive assessment of ABPM were inferred by analytic hierarchy process (AHP). Clinical ABPM data in normal blood pressure group and hypertension group were used for testing the fuzzy measurement system. The score obtained by the system was consistent with subjective clinical assessment of ABPM.

View full abstract

Neural network has attracted interests to develop the knowledge base from the medical database. Some neural networks have been applied to the diagnostic system, but not to the epidemiological analysis. Non-linear characteristics derived from neural network seems to be more appropriate to construct disease models compared to ordinary stastical methods. This study aims to clarify the applicability of neural network to the representtation of the hypertension model. From medical database, 598 cases were chosen to the learning group, and 597 cases were chosen to the testing group, randomly. Input variables used in this study were sex, age, smoking and drinking habits, body mass index, systolic and diastolic blood pressure, total cholesterol, triglyceride, fasten plasma glucose and uric acid. These input variables were used directly and also used after transforming to the fuzzy memberships. The neural network was learned based on the back propagation of error. The diagnostic accuracies were compared among the neural network directly used input variables, the neural network used fuzzy memberships of input variables and logistic regression as the ordinary methods. The neural network used fuzzy membership showed highest diagnostic accuracy than other methods. This neural network was also able to represent the risk factors associated to the occurrences of hypertension.

View full abstract

This paper describes an extraction of landmarks in a roentgenographic cephalogram by using a neural network and a fuzzy template matching. Two kinds of weighted similarity measures are newly proposed for a fuzzy template matching. The rough region where a landmark is supposed to be located is first found by a neural network. The fuzzy template matching is then performed over this region to find out the exact location of its landmark. The landmark called Sella turcica was successfully found in the actual roentgenographic cephalogram within a permissible error for a practical use.

View full abstract

In this paper, we propose the heuristic clustering for a fuzzy modeling. This approach devide an input variable space into some clusters by using an approach following human's experience and intuition. As input variable space is devided into a lot of subspaces, and we suppose that their relations of input variables and output variables are linear. If unit normal vectors of some subspaces are similar, their are unified as a cluster. Then, some clusters are generated. After the heuristic clustering, we set membership functions of premise parts, and values of consequent parts are learned by a steepest decent method. To verify this approach, we show nonlinear function's approximation as a result. Furthermore, as an example of application, we show fuzzy model for the predicting diagnosis of ectopic pregnancy after the pregnancy effect for infertile patients.

View full abstract