This paper proposes a tendon-driven mechanism for a dexterous robot hand with passivity. A human hand has dexterity to grasp various shapes of objects since it can passively change its grasping configurations according to the shape of an object and to the grasping force. Therefore six tendons of a human finger that play major roles to grasp are selected and are simulated by an artificial tendon-driven mechanism. A tendon-driven mechanism and a back-drivable actuator can realize a passive property. That is a finger's attitude is determined by the interaction force with a grasping object's surface and by direction of a grasping force. A spring is inserted between a tendon and a conventional servo motor in order to realize the back drivable actuator. The passive property is mechanically realized so that the finger can respond to change of the grasping conditions without delay. As a result the passive robot hand can improve grasping stability against unexpected external force. The advantage and disadvantage of the proposed mechanism are investigated by some experiments.

View full abstract

This paper proposes a logos drawing system applying Interactive Genetic Algorithms in order to reflect user's logos image. One logo is expressed as a chromosome, which has information on its components' sizes, positions, and colors. The system draws logos by repeating the system generation of logos and user's evaluation of them as follows. The system generates logos in form of chromosomes and presents them to the user. The user evaluates these logos based on his/her logos image. Then the system generates the next generation chromosomes reflecting the user's evaluation. This paper also shows subjects experiments and confirms the validity of the present system.

View full abstract

Recently, many methods of data analyses using soft computing have been proposed. In this paper, we discuss several hybrid methods which represent a fusion of representative methods in multivariate analysis and techniques in soft computing. Also, we discuss the difference between the hybrid methods and the conventional statistical method.

View full abstract

Container loading problems have been studied as optimization problems for packing many boxes on containers by taking account of safety, availability of capacity and easiness of carrying in and out. In this paper, A family of automatic container loading problems is studied and algorithms are proposed. The algorithms are constructed with metaheuristics and include flat or vertical loading schemes, loading efficiency, stability of loaded objects, and computational requirement.Commonsense and expert knowledge incorporation is moreover considered to combine with metaheuristics. Some numerical examples are given.

View full abstract

A control based on human operating knowledge is flexible and good for the nonlinear system. However, it is difficult to incorporate human control into controller. In this paper, we acquire operating knowledge by constructing a model on computer and operating at extended simulation-time and develop an intelligent controller based on human operating knowledge. This development method is applied to the double pendulum which is an unstable nonlinear system. The effectiveness of this method was confirmed by experiments.

View full abstract

The purpose of this paper is to design a mathematical model for a robot vision, that can express states of human's consciousness, especially waking and sleep. Human beings process external information obtained from sensory organs and internal information accumulated within a brain flexibly and in parallel. The proposed mathematical consciousness model, that mimics this human's information processing system, controls a ratio among multiple kinds of external and internal information processing run in parallel. In a waking mode, the external information is mainly processed. In a sleep mode, large proportion of processing is paused, or the internal information is mainly processed.

View full abstract

This paper considers a parameter calculation method in facial caricature drawing with words in order to reflect user's impressions of model face to a facial caricature. A user chooses the most impressive facial feature and parameters values are calculated so as to emphasize the chosen part. This paper also shows the results experiments to confirm the validity of the proposed calculation method.

View full abstract

The idea of a move type ticket gate machine is described.

View full abstract

This paper describes a new approach for tracking circular traffic signs from image sequences. As a part of traffic sign recognition system, the traffic sign tracking will improve the performance of recognition. By tracking the signs, search space is reduced and misdetection caused by temporal occlusion could be avoided. We propose a new blob tracking called two-layer blob tracking to track the signs from frame to the next. This tracking does not require an accurate model, which is essential in the Kalman-Filter tracking. A red color thresholding followed by a blue color thresholding are used to produce the two-layer blobs. To track the blobs between two successive frames, a blob matching using ring-partitioned matching is employed. It could match the rotated, resized and occluded blobs effectively. The proposed tracking method could track the traffic signs from a moving camera, without have to consider some restrictions on the speed and movement of vehicle, and camera installation.

View full abstract

This paper shows the result of the nitrogen oxide estimation modelused in the microscopic road traffic simulator MITRAM in thisresearch. "To decrease the amount of the toxic substances." is a majortask for all human beings. The large number of the cars can beregarded as an exhaust source of the toxic substances emitted to theatmosphere. In the cities, the nitrogen oxide vented by cars bringsphotochemical smog pollution to the urban place. However, in thispaper, the signal control method has been confirmed to be able to easethe traffic congestion in the urban area, which as a result can reducethe emission amount of the nitrogen oxide. Moreover, the correlationbetween the traffic congestion and the amount of the oxide of nitrogenhas also been made clear in this paper.%This paper shows that the oxide of nitrogen estimation model into microscopic road traffic simulator MITRAM in this research. "The number of poisonous substances is decreased" It is a big problem for man. The existence of the car cannot be overlooked as an exhaust source of the poisonous substance that exists in the atmosphere. In the city, the oxide of nitrogen exhausted by the car causes the urban ozone. Then, the signal control method was confirmed to work over measures to ease the traffic jam in the urban area and the relation to the amount of the oxide of nitrogen exhaust was confirmed.

View full abstract

The fuzzy inflow traffic control has been applied on urban expressway to maintain smooth and safe traffic. The advanced inflow traffic controller would be recommended corresponding to the recent progress of traffic control technologies as well as soft-computing. The flexible and robust traffic controller can be created with the formulation of fuzzy logic. Furthermore, the optimization of parameters for efficient traffic controller can be performed by genetic algorithm as well. It would be confirmed that the advanced inflow controller provides the efficient traffic on the urban expressway.

View full abstract

The purpose of this study is the inspection of analysis methods of the configurational characteristics of cities. The main study is comparison and examination with the modified box-count method proposed by Sato and Shingu and the box-count method. Forms of road-networks of six areas in the Tokyo metropolitan region are used as the objects of the analysis.

View full abstract

Congestion problem in the road traffic is a major factor that causesvarious serious social issues such as the environmental contamination,the economic problems, etc.Therefore, corresponding policies are in demand to solve thecongestion problem by means of the maintenance of the road conditionsand the optimization of the traffic signals.But before the policies are established, we need a verification of theeffectiveness of the decisions, where a valid approach is a roadtraffic simulation.In this paper, a road traffic model, which is mainly composed of roadintersections as its major objects, has been proposed for thesimulation of the large-area road traffic.The road model is used in MITRAM, the microscopic road trafficsimulator, which has been well developed by us. And the large-areaurban road traffic is simulated using this model.

View full abstract

Many uncertain factors may be involved in the perceived travel time in the real world. The idea of fuzzy travel time has been introduced to formulate the fuzzy traffic assignment. The fuzzy standard travel time approach has been formulated with possibility measure as well as necessity measure. The algorithm may be applied to real size networks in this study. Impact of Information provision can be evaluated by the traffic flow estimation. As the perceived travel time is influenced by the traffic information, the fuzziness in traffic flow might be discussed under the impact analysis of information provision.

View full abstract

In the human brain, the logical process is performed at the left hemisphere and the intuitional process is performed at the right hemisphere. We recorded event-related potentials (ERPs) by the electroencephalogram (EEG) during subjects were performing selections of an answer. The subjects were asked to add double figures and choose an answer from the two choices. Choices were two types; Correct: the correct answer was included in the choices, and Approximation: the correct answer is not included. We analyzed by using the equivalent current dipole localization (ECDL) to compare the brain activity in the meantime. ECDs were localized to the left middle frontal gyrus (MFG), the left inferior frontal gyrus (IFG) and the left lateral postcentral gyrus (PstCG) between latencies from 1150ms to 1250ms after presentation of choices. Until 1250ms after the PRC, ECDs were localized to the same parts in both type Correct and Approximation. After 1250ms from the PRC, the ERP of the approximation answer had delay from 60 to 100ms for the ERP of the correct answer. In the case of approximation answer only, the ECDs were localized to the right middle frontal gyrus (MFG).

View full abstract

We measured electroencephalograms (EEGs) in recalling long term memories consolidated by mnemonics called the Method of Loci. Then we analyzed brain activities using the equivalent current dipole localization (ECDL). As a result, ECDs were localized to the left supramarginal gyrus (SMG), to the middle temporal gyrus (MTG), called Wernicke's area, and to the left middle frontal gyrus (MFG) called Broca's area. After the Broca's area, ECDs were localized to the prefrontal area (PFA), and then to the MFG and the inferior frontal gyrus (IFG) of both hemispheres. In comparison with the previous fMRI studies, which has a low resolution of a measurement than EEG measurement, by Ichihashi et al., ECDs were localized to the same areas and the results related with transition of brain activities relevant to time sequence were obtained.

View full abstract

A brain mapping in recalling long term memories consolidated by mnemonics called the Method of Loci was carried out using the event related fMRI. The fMRI signals were analyzed by SPM2 and fuzzy c-means (FCM) clustering. The clustering results reveal activations of wide areas from prefrontal cortex to visual areas.

View full abstract

Four types of moving patterns of a circle, started from the center of a CRT, were presented at random order to the subjects. The movement patterns of a circle were downward, upward, rightward and leftward. Electroencephalograms (EEG) were recorded in the meantime. The equivalent current dipole localization (ECDL) was done on each EEG to estimate the source. ECDs were localized to the MT, to the intraparietal sulcus (IPS), to the precentral gyrus (PrCG). Further more, ECDs were localized to the frontal eye field (FEF) and superior colliculus (SC). This tendency is observed for all the subjects and at each directional movement. And the localized ECDs that were localized to the same areas and almost the same latencies after PrCG had the opposite directions between upward and downward movements, and between rightward and leftward movements.

View full abstract

It is thought that the person's state of the brain function can be evaluated by analyzing the utterance voice by a chaos theory method. Our analytical algorithm "SiSECA" shows us that voice signal has not only chaotic behavior but random. Using to measure the level of complexity or disturbance of state of voice signal by SiCECA, it is possible to observe the changes in the mental status of a speaker from the changes in the chaoticity of the speaker's voice. Like the variations in the fist Lyapunov exponents, the variations of CEM are strongly related to the mental and physical status of speakers. This article mainly focus in the aspect of uttered voice chaoticity "disturbance" by aging.

View full abstract

In this paper we have investigated the effect of three kind of facial manipulations on perceived age, in which the shape of face was manipulated using the cardioidal strain, the vertical placements of facial features were manipulated, and the shape was locally distorted. We have discussed the impact of them with comparison of the effects on perception of age using three manipulations.

View full abstract

An interaction with the environment through the body is critical for emergency of intelligence.From the point of view, the researches, in which models of intelligence are pursued by the robot interacted to the real world, are caught on in recent years. We picked out the living neuronal cells from a brain, and make the neurons reconstruct complex networks autonomously on multi-electrode arrays. Using the system we analyzed relationships between features of a neuronal cell and dynamics of a reconstructed neuronal network. We proposed a hypothesis; giving a robot body to this system and carrying out the interaction to environment generate the specific rules in the spatiotemporal pattern of nerve electrical activities, in other word, "intelligence". To elucidate the hypothesis, we reconstructed networks of rat hippocampal neurons on multi-electrode array probe of multi-site recording system for extracellular potential of neurons and connect the living neuronal network to miniature moving robot, Khepera.

View full abstract

This paper proposes a human detection method of a partner robot for natural communication with human and model reproduction for imitative leaning. The proposed method is composed of differential extraction, steady-state genetic algorithm, and a multilayered neural network. The proposed method is applied for a partner robot interacting with a human. Furthermore, we show experimental results about the proposed method.

View full abstract

We measured motion coherence thresholds with and without prior upward motion adaptation. The results showed that the adaptation to fast motion led to increasing upward and decreasing downward coherence thresholds for subsequently presented test stimuli moving faster than the adapting speed. Thresholds for slower test stimuli remained unchanged. Furthermore, the adaptation to slow motion reduced downward coherence thresholds for test stimuli moving slower than the adapting speed. Thresholds for higher test stimuli remained unchanged.

View full abstract

This study executed psychophysical experiments to address the relation between contour integration and figure-ground separation. We obtained results that line segments with ownership attributes can be integrated, and that such integration occurred only when line segments have the same ownership polarity. Those results indicate that the processes of contour integration and figure-ground separation form not unidirectional but bidirectional cascade relation.

View full abstract

This paper presents a general survey on "intelligence" from broad perspectives of living organisms including insects, animals and humans. After the current paradigm shifts commonly occurring in the interdisciplinary academic areas are reviewed, we show that the commonly focused interest therein is reconsideration about the mutual and inseparable relationships between the external environment and the internal of the agent that is an actor, an observer, a cognizer, and an interpreter. We put an emphasis upon a fact that autonomous systems can be characterized by their self-organizing capabilities driven mainly by the internal coherence produced by their internal mutual relations, rather than be described by inputs from an external environment. Then, we introduce the subject of semiotics, a new interdisciplinary branch of science, and its potential contribution to bridging between the biological intelligence and machine intelligence. We also describe about its relationships with the current hot topics of embodiment and symbol grounding often discussed by the robotics people.

View full abstract

TAM (Topographic Attentive Mapping) network is a biologically-motivated neural network with Gabor function type receptive fields. However, the sturucture of receptive fields is a mono-layer, and there is a lack of performance for rotation characteristic of images. In this paper, we formulate a new TAM network structure with extensive and multilayer receptive fields. We also show the usefulness of TAM nework using some examples of character recognitions.

View full abstract

Computer Shogi programs are still in development to defeat the professional players. One of the main problems exists in estimating the game phases. It is said that there are three phases in the Shogi game, so called, opening, middle and endgame phase. The appropriate strategy differs depending on the game phase. In the present paper, we have tried the classification of the phases using piece strength using Self-Organizing Maps, 100 game records of professional players are used to show the feasibility of our approach. All of the games classified the phase into three classes in piece moving order.

View full abstract

This paper introduces a new data visualization method to be used in the analysis of stock change. To visualize stock change effectively, we focus on the change of the continuous session, since the change of the continuous session can be represent the change of investor's minds. In order to visualize continuous session as investor's minds, we apply self-organizing map to classify investor's minds. The results are demonstrated proposed methods can be visualized stock change and inventor's minds.

View full abstract

I suggest a model close to the brain of a human being that finds it easier to recall the collocation used in the nearer past by expressing the learning process of a human being for the way how words link to each other with the transition of weight between them incorporated with the concept of oblivescence by reducing the weight over time. In addition, I generate a chunk that lumps two words together as one that allows me to deal with it as an idiom with a new meaning.Finally, I deploy a system that executes the prediction of input and evaluate the proposed method.

View full abstract

Self-Organizing Maps (SOM) apply neighborhood learning to enable the creation of close output for similar input. This feature is effective for classification in complex problems. On the other hand, normal SOM cannot directly handle time-series data. For this research we investigate methods applying SOM to classify time-series data. We used delayed-time units in order to map time-series data into the input pattern space.The emission of dioxins from waste incinerators is one of the most important environmental problems today. It is known that optimization of waste incinerator controllers is a very difficult problem due to the complex nature of the dynamic environment within the incinerator. For this experiment, we used sensor data from a waste in cinerator plant as the time-series data, and aimed at classifying instances of dioxin emission. Through computer simulation, we showed that the tested SOM method gave good results in classifying low and high dioxin emission combustion states.

View full abstract

SOM(Self-Organizing Map) by T.\ Kohonen is one of methods for classifying automatically data of multi attributes. This method is only applicable to numeric data but not ambiguous data. We expressed ambiguous data as possibility distributions and proposed the SOM that can input possibility distributions. We improve the definition of the distance between two triangular type possibility distributions and compare the methods.

View full abstract

This article introduces our proposal medical systems aided by fuzzy logic. We focus on the reason why we use fuzzy logic in the medical imaging and ultrasound system development. In the contents, we briefly describe medical image processing, and describe the three examples with emphasis on why we use fuzzy logic in the development. Finally, we briefly describe a conclusion and the normal concept of human.

View full abstract

This presentation presents the recent trend of image processing and its application first, and then discuss about intellectual image processing.

View full abstract

This paper describes an image retrieval system aimed at reducing the semantic gap by adopting relevance feedback. To reduce the gap between low-level features and high-level semantics, in this image retrieval system, users'sketches play an important role in relevance feedback. Experimental results show that this method is effective in retrieving some images.

View full abstract

Intelligent scissors is an interactive image segmentation method that allows a user to select piece-wise globally optimal contour segment that correspond to a desired object boundary. However, the intelligent scissors method is too sensitive to a noise or texture patterns on the image. In this paper, we proposed a new intelligent scissors introduced the concept of the separability, in order to improve the extraction ability of object boundary. The effectiveness of the proposed method is confirmed by some computer simulations employing images include weak object boundaries.

View full abstract

In the past few decades, the incidence of malignant melanoma has increased gradually in most parts of the world. Although advanced malignant melanoma is often intractable, early-stage melanoma is curable in many cases. Therefore, early detection and diagnosis of melanoma is essential to the reduction of melanoma-related deaths. We have established a practical internet-based melanoma screening and data collection system. Key components of the system are a new dermatologist-like tumor area extraction algorithm and a neural network classifier. The dermatologist-like tumor area extraction algorithm achieved superior extraction performance and the neural network classifier achieved classification accuracy of 97.3% in sensitivity and 86.1% in specificity with leave-one-out cross-validation test of 319 dermoscopy images. Furthermore, our system supports SSL transaction and requires only several seconds to complete a procedure. Now we are preparing to make our system into commercial viability. Our current issues are advancement of the system generality and development of trade edition of expensive dermoscopy. In this report, we report the current status of our diagnostic system.

View full abstract

Human lung is divided into five distinct anatomic compartments called lobes. Segmenting lung lobes from Multidetector-row computed tomography (MDCT) images can provide useful information for surgical operation and diagnosis of pathology. We have proposed a method for segmenting lungs lobes from MDCT images. The method estimates boundary surface between the lung lobes based on lobar fissure and tubular tissues (blood vessels and bronchi). The method can find the boundary surfaces within the error of 1.5 mm. Almost of errors are caused by fixing nodes outside the lung. This paper proposes a new moving method of nodes in order to improve the lung lobe segmentation algorithm.

View full abstract

An efficient technique for impulsive noise detection and removal for color images is presented in this paper. In order to preserve edge/detail of image, filtering is performed only corrupted pixels. Thus, all noisy pixels must be detected. The pixels detected as noisy are regarded as the missing value. Thus, the interpolation technique can be used for estimating of missing value. In color natural images, all color channels have very similar characteristics such as texture and edge location. In the demosaicking problem of the charge-coupled device (CCD) samples, interpolation of the missing color component is performed using the available another components. We can use the demosaicking methodology to estimating the value of noisy pixels. This paper presents a novel method for restoring impulsive noisy color images, which is constructed by the noise detector and the interpolator based on the demosaicking methodology.

View full abstract

An image processing system for beautifying human face images is proposed using interactive evolutionary computing. Authors proposed before a nonlinear digital filter bank system which removes undesirable skin components, such as wrinkles and spots, from human face images to make the face look beautified. The system was so effective, but it requires a lot of parameters to be adjusted, which were set by trial and error. In this paper, the evolutionary computing is newly introduced into the nonlinear digital filter bank to set the values of the parameters so that the output face image be highly evaluated on the basis of subjective criteria. The new face image processing system is verified to be effective in processing some actual face images.

View full abstract

For the estimation of an out-of-focus PSF, it is effective to obtain the radius of shade ring which appears on the amplitude spectrum of an observed image. However, the estimation is very difficult in case where the observed image is not so blurred and corrupted by an additive noise. Because the shade ring tends to get lost in high frequency region dominated by low intensity. This paper proposes a robust PSF estimation method based on the concept of vector convergence. The vector convergence focuses attention only on the vector direction and extracts contrast invariant features. The effectiveness of the proposed method is confirmed by some computer simulations.

View full abstract

In this paper, an approach to reconstruct 3D shape of the buildings by using binocular stereo method is proposed. Firstly, the 3D distance of outlines for buildings is estimated. The binocular stereo method has a problem that not keeping optical axes parallel leads to serious error. Therefore, the camera model considering camera rotation is proposed in order to revise the error. Then the planes constructing the building are extracted from the combination of outlines of buildings, and the shape of building is estimated from these planes. To show the efficiency of the proposed method, the reconstruction experiment is performed for both the streetscape images of miniature models and the buildings. Consequently, the extraction of planes from miniature models was achieved. Further, the 3D distance of outlines for buildings was successfully estimated, and the efficiency of the proposal camera model was confirmed.

View full abstract

In this paper, we proposed a new framework to perceive the relative depth information from a 2D scene using scale variation of texture image. To extract feature vectors from textured patterns, higher order local autocorrelation functions are utilized at each scale step. The hierarchical linear discriminant analysis is employed to classify the scale rate of the feature vector, which can be divided into subspaces by recursively grouping the overlapped classes. In the experiment, relative depth perception of 2D natural scenes is performed on the proposed method and it is expected to play an important role in natural scene analysis.

View full abstract

In this paper, a novel fuzzy-clustering-based-approach for object recognition is proposed. In situations involving multiple objects that can be replaced by a primitive model, the proposed method can be applied without prior information about the number and the shapes of the objects. The approach is composed of three stages. In the first stage, 3D data is reconstructed using stereo matching from a stereo image that includes multiple objects. Next, the 3D data is separated into objects by using a Fuzzy c-Means algorithm augmented with a criterion about the number of clusters. Finally, the shape of each object is extracted by Fuzzy c-Varieties with noise clustering. The effectiveness and validity of the proposed method was shown using both preliminary simulation data and real data obtained from stereo matching.

View full abstract

A support implant has been used to reconstruct an artificial hip joint on the acetabulum in revision total hip arthroplasty (THA). After revision THA, we should diagnose periodically form of the support implant because the support implant may be distorted or broken. This paper proposes an in vivo computer-aided diagnosis (CAD) system that can evaluate the form of the support implant using a finite element method (FEM) and a genetic algorithm (GA). The proposed method searches for optimum local deformed models using a GA from a local deformed model database. The direction and the strength of the stress for the support implant are required from optimum local deformed models. As a result of applying to simulated MDCT images of horseshoe-shaped phantoms, the proposed method could recognize the correct local deformed models successfully. Additionally, We could obtain the direction and the strength of the stress for horseshoe-shaped phantoms.

View full abstract

An efficient visual tracking algorithm based on particle filtering is proposed. To reduce the time complexity, a procedure for preventing redundant image processings used to find a target object in an image sequence is introduced. This reduction is achieved by determining equivalence samples (called particles), which represent a state distribution, in the digital space. To evaluate the performance of the proposed algorithm, a noisy image sequence,consisting of 90 binary images with 320*240 pixels, is used.The result shows that the execution time decreases with increasing the number of particles and reduces by 32.9% compared with the conventional way when particles are 10000 in number.

View full abstract

This paper proposes an evaluation rule by analyzing a building form and the evaluation of Kansei, and verifies that precision.

View full abstract

In various products such as the dresses and the interiors, special knowledge of not easiness but the color is demanded from doing the arrangement of color where the entire harmony was able to be taken while we are doing those arrangement of colors. In this research, it proposes the system that makes a fuzzy rule and the membership function made based on the color harmony theory to support such an arrangement of color, judges color harmony automatically, and makes the arrangement of color.

View full abstract

In this paper, analysis of questionnaire survey to managers on food processing is introduced. Management of losses due to out of stock and surplus stock in a retail food store is well known as an important element in freshness food management. The manager who has a responsibility in side dishes area faces to difficult decision making on food processing such as quantitative and qualitative arrangement. According to the analysis, their knowledge are extracted by a techniqueof text mining.

View full abstract

In this paper, we propose a method for investigation of Kansei information with Fuzzy C4.5 decision tree. In the case of Kansei information investigation, finding rules from constructed decision tree is required to be simple and easy to be understood. In our previous research, we had investigated the change of structure of constructed Fuzzy C4.5 decision tree from various types of membership functions and parameters for pruning. We found some optimized decision trees and rules about usability of golf clubs, which was an example for Kansei information investigation. In this paper, we evaluate stability of constructed Fuzzy C4.5 decision tree. We also compare the estimation abilities of fuzzy C4.5 decision tree with standard C4.5 decision tree.

View full abstract

This article presents user model for multimedia contents delivery service related to new network routing control. As an application of new network routing control, we have studied about contents management system based on Kansei information processing. The purpose of this study is to develop techniques which enable to provide contents and service based on user's subjective requests using best network routing. In this paper, we report the study about user model by using data mining technique based on the investigation of tolerance against network troubles for multimedia contents delivery service.

View full abstract