Monitoring and evaluation (M&E) are an essential public management tool that can be used to improve the way governments and organizations achieve results. Just as governments need financial, human resource, and accountability systems, governments also need good performance feedback systems. The introduction of a results-based M&E system takes decision- makers one step further in assessing whether and how goals are being achieved over time. The objective of this paper is to develop a new approach to treating rough sets. The method will play an essential role and will be employed in the evaluation system. This research explores new possibilities for the application of decision support systems. A prototype Urban Innovators Systems has been developed to support this collaborative model of public participation. It provides an interactive problem-solving platform to urban designers by implementing the proposed three major functions: Creative City assessment, Sim-Architecture and a best practice database in the prototype Urban Innovators Systems

View full abstract

This paper studies reliability optimization of a series-parallel system with fuzzy random lifetimes. A fuzzy random reliability model is developed to maximize the system reliability. Since the reliability function consists of fuzzy random parameters, classical mathematical programming methods is not applicable to the reliability model. Therefore, in order to solve the model, a fuzzy random simulation method is first proposed to compute the system reliability. Furthermore, a hybrid binary particle swarm optimization (BPSO) algorithm incorporating the fuzzy random simulation is proposed. Finally, a numerical example is provided to illustrate the proposed hybrid algorithm.

View full abstract

This paper describes a basic system to detect humans around a vehicle through fisheye video cameras. In order to monitor the surrounding environment of a vehicle, a camera with wide field of view is needed, so we first choose a single fisheye camera to take videos. To deal with the video captured by the camera on a static vehicle, background subtraction method is employed first. The background subtraction is the process of separating out foreground objects from the background in a sequence of video frames. In this paper, we compare three different algorithms of background subtraction to preprocess the images and try to find out the proper algorithm to be applied in our human detection system.

View full abstract

In real option pricing, it is impractical to assume the net present value of expected cash flow payoff as an exact number because it is a forecasted vague one. The net present value can defined a fuzzy number to express its estimated uncertain values and the Black-Scholes formula is used to price a real option. A modified pricing approach to real options is thus proposed to transform the forecasted uncertain values evaluated by experts into some normal fuzzy numbers. The distance between fuzzy variables are introduced to construct weighted vectors. It is verified rational to estimate the net present value of expected cash flow payoff with a normal fuzzy number is. A numerical example is given to illustrate the validity of the proposed approach .

View full abstract

The objective of this paper is to build a regression model with confidence intervals where such confidence intervals are expressed through the expectations and variances of fuzzy random variables. First, a general regression model for fuzzy random data is introduced. Then, using expectations and variances of fuzzy random variables, a regression model with confidence intervals is established. Given the nature of such modeling, we will be referring to these models as fuzzy random regression models with confidence intervals. The proposed regression model gives rise to non-linear programming consisting of the productions between involving fuzzy numbers or interval numbers. The inherent non-linearity of the optimization makes it hard to exploit techniques of linear programming and we need to resort ourselves to some heuristics. Lastly, an illustrative example is provided.

View full abstract

Genetic fuzzy rule selection can obtain a simple and accurate fuzzy rule set with linguistically interpretable rules from numerical data. In the first phase of our method, a large number of fuzzy rules are extracted from numerical data by a data mining technique. In the second phase, a small number of the extracted rules are selected by a genetic algorithm. When we apply this method to large data, each solution needs much longer time for evaluation. To cope with this problem, we proposed the parallel distributed implementation of genetic fuzzy rule selection. In our proposed method, not only population but also training patterns are divided into subgroups. Each subgroup is assigned to one CPU. The proposed method can drastically reduce the computational time without severer deterioration of classification performance. In this paper, we examine the effects of data subdivision on the classification performance and computational time.

View full abstract

We compare two feature selection methods, logistic analysis stepwise procedure and decision tree. By comparing these selection methods, we discuss rating scale for feature selection. We use credit card transaction data for comparison experiment.

View full abstract

This paper provides a new evolutionary multiobjective optimization method for automatically optimizing the network structure of recurrent neural networks. The feature of the proposed method is that it involves a procedure of intensively exploring a region including solutions with small training errors in the Pareto frontier, instead of finding a whole set of the Pareto optimal solutions. Several numerical experiments are executed in order to show the advantage of the proposed method over the existing effective algorithm by Delgado et al. with respect to the generalization capability.

View full abstract

The privacy-preserving recommendation system enables us to evaluate the recommended value without leaking the private information of users to service providers. The large overhead in performing cryptographic operations in proportion to the number of users and the number of items is the current issue. In this article, we propose an efficient scheme by reducing the dimension user evaluation matrix in the sets of items and users.

View full abstract

The clonal selection principle is used to explain the basic features of an adaptive immune response to a antigenic stimulus. It established the idea that only those cells that recognize the antigens are selected to proliferate and dierentiate. This paper explains a computational implementation of the clonal selection principle that explicitly takes into account the anity maturation of the immune response. The clonal selection algorithm by incorporating receptor editing method, RECSA, has been proposed by Gao. This paper introduce the immune memory cells into RECSA model. In order to verify the eectivness of proposed method, we try to classify MONK's problems and the medical database of Coronary Heart Disease (CHD) database.

View full abstract

Owing to the progress of RFID readers and RFID tags, we can follow actions of customers in a shop. We can anticipate that the analysis of the actions helps the decision makers of the shop operate the daily business. This paper deals with the RFID data collected from an apparel shop and proposes the analysis method of the data. Also, the paper verifies the effectiveness of the method using the data collected in real world environment.

View full abstract

Most matter in nature and technology is composed of crystals and crystal grains. A full understanding of the inherent symmetry is vital. A new interactive software tool is demonstrated, that visualizes 3D space group symmetries. The software computes with Clifford (geometric) algebra. The space group visualizer (SGV) is a script for the open source visual CLUCalc, which fully supports geometric algebra computation. In our presentation we will first give some insights into the geometric algebra description of space groups. The symmetry generation data are stored in an XML file, which is read by a special CLUScript in order to generate the visualization. Then we will use the Space Group Visualizer to demonstrate space group selection and give a short interactive computer graphics presentation on how reflections combine to generate all 230 three-dimensional space groups.

View full abstract

The system in which the three-dimensional morphological structure of a neuron is reconstructed from the sequence of two-dimensional images obtained by a confocal laser scanning microscope has been developed. To investigate the structure and the characteristic of a neuron, first, we extract a neuron shape from the sequence of two-dimensional images. Next, the three-dimensional morphological structure must be constituted by use of an extraction result. Since these operations have so far been executed manually, large-scale time and work have been wasted. In contrast, in proposed system, the branching structure and diameter of dendrites are extracted by the single-seed distance transformation method, and compartment neuron models are reconstructed by using the obtained results. In addition, the operation procedure is simplified by use of the graphical user interface. Therefore, the proposed system can support a user intuitively and effectively.

View full abstract

Recently, companies often carry out questionnaire(s) and develop marketing strategies. There are usually two types of forms for the answer of a questionnaire. One is the form to select prepared answers and the other is free text form. The true message might be in the text form rather than the numerical part, then the analysis of free text form is needed. The amount of text in a questionnaire is, however, usually large and difficult to read whole text data for analysis. This study tries to develop a free text analysis support system which visualizes relationships among respondents based on their texts and shows their opinions using graph structure of keywords. First, this paper proposes the extraction method of important keywords in their opinions based on the modification relationships. Next, it clusters the respondents interactively on visible space by MDS. Finally, it shows their opinions using HK Graph which can visualize the relationship among words with hierarchical network structure based on the co-occurrence information for the keyword graph.

View full abstract

For more efficient machine learning, it is important for a lot of machine learning methods to construct hierarchical model. Then, it is popular approach to use linear method such as PCA or MDA to visualize the data distribution in the case of high-dimensional data. However, those linear methods are not appropriate for the construction of hierarchical model based on a dis-similarity (distance) between instances, because they do not consider dis-similarity between instances. This paper proposes two-dimensional dendrogram map based on multi-dimensional scaling (MDS) method using spring model which considers dis-similarity between instances. This paper studies an interactive construction of hierarchical structure for machine learning based on the visualization result. In the experimental result, it was confirmed that the proposed method was effective for feedback into the machine learning.

View full abstract

We propose a fuzzy logistic regression model for Electroencephalogram Data. The purpose is to discriminate between kinds of human thinking through the use of Electroencephalogram signals. In our proposed method, we introduce classification structure obtained by fuzzy clustering into the logistic regression model in order to improve the precision of the discrimination. Several numerical examples are demonstrated to show a better performance of our proposed method.

View full abstract

Reinforcement learning has been studied as an unsupervised learning method. Alternatively, BCI(Brain Computer Interface) comes into the research limelight. However, nonsynchronous spontaneous action potentials and evoked action potentials exist contain brain signal, and we need an interface model which exists between brain and machine for control and stability. In this paper, we propose a collaborative learning system consisting of reinforcement learning and brain signal in BCI. Brain signal is interpreted as a deliberate assignment of the subject, and we utilize reinforcement learning in control and stability for BCI. We apply the collaborative learning to maze problem and show the usefulness of the proposed system.

View full abstract

This paper presents feature selection of electroencephalogram (EEG) signals and image classification based on machine learning. First, to classify two kinds of image, we discuss a way to acquire frequency feature from EEG signals. Next, a support vector machine (SVM) is employed to image classification based on the frequency feature of EEG signals. The preliminary experimental result shows that the classification result via SVM is superior to those via discriminant analysis, neural networks and partial least squares regression. In addition, we propose an innovative algorithm, which combines the SVM and the so-called Adaboosting, to enhance classification rate for EEG signals and to map a brain-activity for EEG signals. As a result of applying the proposed technique to wheelchair control experiments, the experimental results demonstrate the utility of the SVM algorithm with Adaboosting and the viability of accomplishing wheelchair control only by EEG signals.

View full abstract

Recently, a lot of researches on Brain Computer Interfaces (BCIs) have been reported. It is expected that BCIs will help patients like those with Amyotrophic Lateral Sclerosis (ALS) to control a wheel chair or to communicate with other people just by thoughts. BCIs recognize thoughts through analysis of biological signals such as Electroencephalogram (EEG), and those associated with motor imagery have been investigated. Bereitschschafts potential (BP) and event-related desynchronization (ERD) are well known as features of EEG related to movement, and EEGs of different movements have different temporal patterns. The authors have proposed a method to extract such features and have shown that the proposed method is effective for EEGs of actual movements. In this paper the proposed method was applied to EEGs of motor imagery, and a classification accuracy for test data set was 93%.

View full abstract

A Brain-Computer Interface (BCI) is a system which could enable patients like those with Amyotrophic Lateral Sclerosis to control some equipment or to communicate with other people, and has been anticipated to be achieved. One of the problems in BCI research is a trade-off between speed and accuracy, and it is practically important to adjust those two performance measures effectively. So far the authors have considered BCIs as communications between users and computers, and have proposed Reliability-Based Automatic Repeat reQuest (RB-ARQ), an error control method. It has been shown that, with Linear Discriminant Analysis (LDA) as a classifier, RB-ARQ is more effective than other error control methods. In this paper, Support Vector Machines (SVMs), one of the most popular classifiers, is applied to RB-ARQ. Also, a quantitative comparison among the error control methods has been done for the first time. The performance of BCIs have been further improved by RB-ARQ compared to the top ranked methods in the BCI competition with no significant difference between LDA and SVM.

View full abstract

The authors have recorded EEGs from subjects reading silently four types of arrows presented on CRT. Each of arrows has directions of upward, downward, leftward or rightward. The canonical discriminant analysis is applied to the single trial EEGs. Four channels of EEGs at the right frontal and temporal had been used to discriminate the four types of symbols. They are 4, 6, 12 and 14 channels according to the International 10-20 system. Sampling EEGs were taken form 400ms to 900ms at 25ms intervals. So the numbers of variates are twenty-one by four, i.e. eighty four. Results of the discriminant rates by use of the present system were more than 90 %. These results are improved from the precedent our research. By four type code of infrared rays according to the discrimination results from a PC, the authors have controlled a micro robot the e-puck.

View full abstract

The "Morphopsychologie", by Dr. Louis Corman is analysis for relationship between personality and physical features of face implicitly. The aim of this research is to construct a quantitative model of physical feature of face and personality, and discuss the results of research based on "Morphopsychologie". For this aim subjects were completed a 240 item self-report personality questionnaire (NEO-PI-R test) and each personality was assessed based on the big five personality dimensions and classified into high and low groups respectively. Then, facial characteristics of subjects who were considered to be higher or lower in five personality dimensions, are analyzed.

View full abstract

Most of conventional associative memory models suppose that patterns to be stored are distributed over a pattern space such that each pattern is located at almost the same distance from all the others. In other words, things (concepts) are assumed to be encoded into patterns which do not preserve similarities (correlations) among them. However, this assumption seems implausible, since our actions (e.g. recognition, thinking, and learning) are often based on similarities of concepts. This paper supposes that concepts are represented as hierarchically correlated patterns in which concepts of the same category are encoded into similar patterns, and then proposes a model which successively retrieves patterns belonging to the same category.

View full abstract

Communications between human and the robot have been widely researched. The robot enters not only industry but also the home, and helps people's daily lives or amuses. Various robots have been developed to communicate with human. However, there is still a problem that the natural communications cannot be done. For one thing, the robot is not recognizing the person who is speaking. Therefore when talking, the content of the utterance does not be suitable for the place, and the flow of the conversation is not understood. To solve this, it is necessary to choose the content of the conversation suited for each person. The purpose of this research is to recognize human's faces for the natural communication. The method of recognizing human's face is clarified based on study results of the cognitive psychology, and it proposes the methodology concerning the feature extraction to detect human's face.

View full abstract

Recently, BCI(Brain computer interface) has been come into the research limelight. We have been investigating action potentials of rat hippocampal neurons cultured on a dish with 64 micro planer electrodes. However, we don't exactly comprehend logicality in micro planer electrodes. In this paper, we analyze logicality in three electrodes using fuzzy operator consisting of t-norm operator and t-conorm operator, and discuss logicality in cultured neuronal network.

View full abstract

Galvanic Vestibular Stimulation (GVS) is known as a method of giving influence on sense of equilibrium. The vestibular system is stimulated by a weak current through an electrode placed on the mastoid behind the ear. In this paper, we had a motion guidance experiment of walking and the standstill operation by using GVS to evaluate the effect. And we measured the amount of the motion guidance with a motion capture system. We propose a motion guidance model to the change in a fixed fuel spray travel, time, the walking speed and compared it with the experiment result.

View full abstract

A precise tissue characterization of coronary plaque based on fractal analysis by using radiofrequency (RF) signal is proposed. The RF signal is obtained by intravascular ultrasound (IVUS) method, which is used for the diagnosis of the acute coronary syndromes (ACS). In the proposed method, the RF signals reflected from the different tissues are supposed to have the different complexities. The effectiveness of the proposed method is verified by a series of experiments using IVUS RF signals obtained from the coronary arteries of a rabbit and a patient.

View full abstract

In a diagnosis of the acute coronary syndrome, the tissue characterization of coronary plaque is very important. The integrated backscatter (IB) analysis is a typical method for the tissue characterization by using the intravascular ultrasound method. The IB analysis employs the IB value, which is a locally averaged power of the backscattered ultrasound signal (RF signal). However, the IB analysis is not always accurate for the classification of the tissues. We propose a novel method for the tissue characterization by using a sparse coding. In the proposed method, the RF signal is coded by the sparse coding, and the code patterns are used for classification. Experiments are performed to classify the tissues of the plaque into the fibrous and the lipid tissues using the RF signal, which is obtained from a rabbit coronary artery. The effectiveness of the proposed method has been verified by comparing it with the IB analysis.

View full abstract

This paper proposes distance-type TS reasoning method, in which the consequent part of fuzzy rule consists of a function. Moreover, learning algorithm is proposed for the distance-type TS reasoning method and applied to a diabetes diagnostic system, showing the satisfactory learning results.

View full abstract

As an extension of TS fuzzy reasoning method, "weighted TS fuzzy reasoning method" was proposed by us in which the consequent part of fuzzy rule consists of a function with an weighting function. This paper proposes learning algorithm for the weighted TS fuzzy reasoning method by using steepest decent method, and compares learning results with the cases by simplified reasoning method and TS fuzzy reasoning method by applying to a diabetes diagnostic system.

View full abstract

It is natural to think that couples who work at a company or a person who lives by her/himself want to cook food for themselves as quickly and easily as possible when they are busy. However, to keep having the same food they can easily cook fed them up, therefore, it should be preferable for them to be recommended a variety of food that they can cook "easily". Currently, there are so many Web sites for cooking recipes, and there are also recipes regarded as "easy" to cook on the sites. However, those recipes are not estimated as "easy" by taking user's conditions into account. The meaning of the word, "easy", would be differently interpreted by each user's conditions. Therefore, in this study, we aim to propose a method to recommend "easy" cooking recipes by considering user's conditions and then develop a recommendation system with the proposed method.

View full abstract

This paper investigates the influence of electronic word-of-mouth information provided by other consumers, called consumer generated information, to a consumer received the information in the context of consumers choosing to purchase products. An effect of increasing the preference of the product, called product attitude effect, and an effect of increasing the expectation of the future with the product, called purchase attitude effect, are introduced to investigate the influence. The experiment design used consists of two steps: (1) subjects choose a digital camera from two by using information from the makers and (2) the subjects change their attitude toward the chosen digital camera by using presented consumer generated information. The significance of these findings is discussed from the viewpoint of disparity in classes of product's characteristics.

View full abstract

This paper considers a method for generating visual summary from discussion log on BBS. We have been developing KGBBS (KeyGraph-based BBS), which enables online discussion while sharing data visualized by KeyGraph. In order to utilize the discussion log of KGBBS, the concept of visual summary is proposed for visualizing the transition of topics during discussion. The visual summary is generated by extracting a comment chain diagram that shows the structure of discussion and visualizing topical transition on KeyGraph. This paper also discusses the concept of visual summary for ordinary text-based BBS by extending the KeyGraph-based visual summary.

View full abstract

Financial analysis is considered a process with clear goals and viewpoints, but in fact, it is often exploring interactive process. So, we need to design interactive Web system suitable for exploratory financial data analysis. We have built a Web based system supporting financial analysis and have observed interaction processes with the system. In this paper, we would report our findings and discuss desirable interaction design for financial analysis.

View full abstract

There are many services in the internet by appearance of web 2.0. Users don't have many ways to obtain useful information for themselves, while there are huge quantities of information. To assist that users find useful information, recommendation system is existing. However, current recommendation system is not considering user's interest. Therefore, we focus attention on the user's interest range. If user's interest range could be identified, it could expect that it is able to recommend more useful information for user. To identify user's interest range, we used the data of the Open Directory Project (ODP). In this thesis, we propose the method of identifying the user's interest range, and discuss usefulness of our method.

View full abstract

The city model consists of one city center, a public transportation facility, and built areas. Let a unit that constitutes the built area be a built-area cell. It is assumed that a built-area cell changes to the Propriety of urban density according to the density of the block where it exists. Urban formation is influenced by the following two factors if it does so. One is the convenience of a public transportation facility and another is the propriety of urban density of a block. Regarding the route of a public transportation facility, two kinds, a radiate pattern and the pattern extended to alignment, are set up, and several sorts of move speeds are set up. And regarding the propriety of urban density of a block, the urban density corresponding to the convenience of public traffic is assumed. A function is set up that propriety of urban density also becomes higher as traffic convenience becomes better. The simulation of the urban form is carried out under the above conditions.

View full abstract

In the planning stage of the building design, the design images of each other may not be understood sometimes to the partner by the differences of sense of value and knowledge between designer and the client. So, if there is a tool to output visually so that a partner can understand the building image easily, it is effective. For the first step of the tool development, the purpose of this study is to investigate a relationship between a specific sensitivity image and the form element of the facade of the office building, for the facade of an office building that is regarded as important factor to express the brand image of the company. Rough Sets Theory was used for feature analysis. The form elements which express the "modern" feeling are as follows; "A same pattern continues from the lower layer to the upper layer" and "The window is full-height glazing" , "The finish of facade is metallic" , etc.

View full abstract

As the landscape act is established, beautiful landscape of urban area is expected. Therefore, the evaluation of urban landscape is an important subject concerning with construction of new building in urban area. The study aims at developing the expert system of urban landscape evaluation to accumulate and effectively apply the expertise knowledge in actual results of landscape evaluation. The pilot expert system for evaluation of urban landscape has been already proposed based on the examples of landscape advices in local city. In the study, the inference mechanism is proposed to handle fuzzy knowledge base for realizing the urban landscape with fuzziness in landscape evaluation. The fuzzy expert system is developed with implementing multi-stage singleton fuzzy reasoning. Since the proposed system is compared to the existing landscape evaluation system with practical examples of urban landscape evaluation, the effectiveness can be summarized practically.

View full abstract

The information provision is an important technique to produce the efficient traffic flow on the network with accidents. This study aims at proposing the practical algorithm for information provision method on the network for efficient traffic management with accidents. It is assumed that the information of traffic accident is provided through the traffic information board system. The simple process to determine the effective pattern of information provision is summarized as practical algorithm for information provision. Therefore, the effective patterns of information provision would be proposed empirically with fuzzy reasoning formulation from the case studies for the urban network of Gifu city. Based on the analysis with dynamic flow, it can be concluded that the fuzzy logic based algorithm can be commonly applicable for practical information provision on urban network.

View full abstract

Recently, technology for producing 3D model space from optical satellite is paid to attention as a technique that takes the place of photogrammetry. But it is known to lack stability because of current photogrammetry. Then, this research tried the production of 3D model space with optical satellites.

View full abstract

The number of deteriorated infrastructure has increased dramatically in many countries. As a means of inspection for existing concrete structures, visual inspections are generally carried out to draw a sketch of cracks. The crack sketching takes an incredibly long time. Therefore, there is a significant demand for advanced tools for visual inspection. Recently, digital image processing technologies have been successfully applied in different industries. As the results, there are a lot of processing techniques. However, it is difficult to find an optimal image processing procedure for extracting cracks from a concrete surface image due to the fact that a broad array of procedures needs to be evaluated. This paper describes a novel approach for automatically constructing the image processing procedure for extracting concrete cracks. A characteristic feature of the study is the adoption of evolutionary algorithms to determine an optimal or suboptimal image processing procedure.

View full abstract

In this study, we focused on knowledge workers' working behavior in workplace. With spatial temporal trajectories got by UWB sensor network, the objective data of knowledge worker s' behavior about when, where and how long was obtained. Then, the working behavior characters influenced by workplace layout were clarified. It was clear that the workers spent most of time in self-desk area and meeting area and stayed longer than 10 minutes per time there. Although workers spent fewer time in support area, they visited there more frequently. It was clear that the knowledge workers always chose the table in meeting space which is nearest from their self-desk for staying.

View full abstract

It is important to know resident's daily basic behavior such as room usage in house design. However, it is difficult to observe the daily basic behavior without seriously invading his/her privacy by common investigation method. Such motivation made us observe when and in which room an elderly living alone was staying in a detached house for a week period using active Radio Frequency Identification (active RFID). Then using Bayesian Network model, we found some behavioral characteristics that showed how she stayed in each specific room and how she moved among rooms. In the result, the resident was inclined to stay longer in the living room than in other rooms. She often stayed in the washing room longer in the afternoon, and stay in the living room longer in the evening. She was inclined to move toward the kitchen and washing room from other rooms in the morning and move to the living room in the afternoon. A connection based on her movement between the living room and kitchen became stronger from morning to the evening. We showed an availability of method to acquire room-using behavior objectively with noninvasive observation technique and to analyze behavioral characteristics by employing Bayesian Network model.

View full abstract

In order to avoid the difficulty of classification of huge number of images, Tagged Image management system based on Lattice Structure (TaILaS) is proposed based on formal concept analysis. TaILaS constructs a concept lattice from a given cross table between objects and attributes, and then the Hasse diagram of this lattice is visualized, where the all attributes are given as tags information. TaILaS can assign each image with multiple categories based on lattice structure, therefore, it is easy for users to treat with huge number of images and avoid the difficulty of their classification. The proposed system is running on an ordinal computer and it is implemented in Processing, a JAVA based programming language. Through visualization experiment using the Corel Image Galley, we confirm that the proposed system is more effective than the existing file management system by tree structure.

View full abstract

A problem of failure in tracking under out of field of view situation has been considered. We propose a method being able to continue the tracking by taking into account geographic information estimated from the background. We deal with a scene that Unmanned Aerial Vehicle is tracking a car. The tracking sequentially estimates the velocity and posture of the car from an image sequence using particle filter. The geographic information is road. The road area is also estimated using particle filter. Unknown road are extrapolated from known road with cubic splines. Target disappearing from the frame is predicted by state space model which imposes prior knowledge that the car normally moves on the road. Effectiveness of proposed method has been shown through tracking experiments with simulation image sequences.

View full abstract

This paper proposes a biometric personal authentication based on the pressure distribution while one step walking. We extract one step from a walk on mat type load distribution sensor and use it to personal authentication. In this method, features with respect to weight movement and foot shape during walking are employed, and then a classifier is developed aided by of fuzzy inference. Those classifiers trained by artificial immune system. We employed 30 volunteers. For each volunteer, we obtained walk data six times. Then, we evaluated this method by five training data and one test data. The experiment result shows 15.0% EER (Equal Error Rate) and 7.0% FRR (False Rejection Rate) in verification (1:1 collation) and identification (1:N collation), respectively.

View full abstract

Car driver's motion estimation is important for developing safety driving systems. We propose to use particle filters for this purpose. Several sensors, such as camera to capture driver's face, velocity sensor using Doppler Effect of microwave for driver's foot motions, and microphones to record and to identify driver's voice, are placed at experimental cabin of driving simulator with three display panels. From these sensor signals, particle filters estimate the information to be obtained based on some elaborated state space models. So obtained information are gathered to server machine for making global recognition of the driver's body motion. We have implemented the data gathering system based on server-client framework on TCP/IP.

View full abstract

Augmented Reality (AR) is an emerging technology that synthesizes the real vision with virtual reality (VR) objects by aligning the objects onto the real world at high precision. AR literature has been growing radically in past few years. As for hardware, for example, cellular phones and PDA also could run state of the art AR system thanks to the recent development. AR becomes a promising tool in various applications including entertainment, surgery support, and other many business works, and it is thought that is about to break to our daily scenes. In this talk, we look through the recent progress of AR and pick up some outstanding papers, and then we introduce our contribution on free viewpoint video generation and visual support for drivers and pedestrians.

View full abstract

Traveling on foot is an intuitive and natural practice in the real world. However, the problem of moving around in the virtual environment on foot is one of the major obstacles to be tackled in virtual reality research. We usually explore a virtual environment using a hand-held controller, even if walking is the most natural locomotion method for human beings. In order to realize sensation of walking in a virtual environment, constructing the same environment in the real world is the best method. However, it is difficult to construct such environment since it costs too much. Therefore, many locomotion interfaces that can give a user the sense that he/she is walking while his/her actual position remains localized in the real world have been developed. In this talk, I will introduce their basic configurations and some applications of locomotion interfaces such as a sharing sense of walking environment and gait rehabilitation system.

View full abstract

Total knee arthroplasty (TKA) is an orthopedic surgery for the damaged knee joint. The movement function of the implanted knee joint in vivo has many questions. Therefore, it is important to analyze the knee kinematics in a diagnosis. There are some conventional methods for analyzing kinematics of the implanted knee joint using X-ray fluoroscopic images and 3-D geometrical models of knee implant. However, almost studies were designed for static images, and there are few studies that considers the continuous movement of knee joint in dynamic image. Thus, the estimation results tend to discontinuous between the neighboring frames. This paper proposes an analysis method of the implanted knee kinematics using digital radiography images with kalman filter. Kalman filter analyzes continuous knee movement and improves estimation accuracy. The experimental results showed that the implanted knee joint angles were estimated by the proposed method with a mean estimation error of 0.31 deg.

View full abstract