We propose a new practical input interface for the recognition of Non-Audible Murmur (NAM), which is defined as articulated respiratory sound without vocal-fold vibration transmitted through the soft tissues of the head. We developed a microphone attachment, which adheres to the skin, by applying the principle of a medical stethoscope, found the ideal position for sampling flesh-conducted NAM sound vibration and retrained an acoustic model with NAM samples. Then using the Julius Japanese Dictation Toolkit, we tested the feasibility of using this method in place of an external microphone for analyzing air-conducted voice sound.

This paper presents a new type of wearable teleoperation system that can be applied to the control of a humanoid robot. The proposed system has self-contained computing hardware with a stereo head-mounted display, a microphone, a set of headphones, and a wireless LAN. It also has a mechanism that tracks arm and head motion by using several types of sensors that detect the motion data of an operator, along with a simple force reflection mechanism that uses vibration motors at appropriate joints. For remote tasks, we use intelligent self-sensory feedback and autonomous behavior, such as automatic grasping and obstacle avoidance in a slave robot, and we feed the information back to an operator through a multimodal communication channel. Through this teleoperation system, we successfully demonstrate several teleoperative tasks, including object manipulation and mobile platform control of a humanoid robot.

Face motion is composed of rigid motion and non-rigid motion. The rigid motion occurs from movements of the human head and the non-rigid motion derives from human's facial expression. In this paper, we present a technique for estimating these rigid/non-rigid motions of the human face simultaneously. First, we test whether the face motion is rigid. If it is rigid motion, we estimate the translation and rotation parameters over image sequences. Otherwise, the non-rigid motion parameters based on the spring-mass-damper (SMD) model are estimated using optical flow. We separate the rigid motion parameters explicitly from the non-rigid parameters for parameters de-coupling, so that we can achieve the face motion estimation more accurately and more efficiently. We will describe the details of our methods and show their efficacy with experiments.

We propose a shape resolution control method applying a tolerance caused by movement to object's shape and texture in order to represent efficiently a textured object that has a detailed structure. It is generally difficult to perceive the error of shape or texture of the object that is moving. Our method applies this error as a tolerance. The efficient object's representation is realized by the shape resolution control that tolerates errors of contour shape and textured surface by the tolerance caused by movement and reduces object's data. It was shown better experimental results of processing time and of the quality of images in comparison with other methods. Thus, it was proved that the method applying the tolerance caused by movement to the object's shape and texture is effective in the representation of textured object that has a detailed structure.

We present a depth video enhancement algorithm in order to provide high quality haptic interaction. As the telecommunication technology emerges rapidly, the depth image-based haptic interaction is becoming viable for broadcasting applications. Since a real depth map usually contains discrete and rugged noise, its haptic interaction produces the distorted force feedback. To resolve these problems, we propose a two-step refinement and adaptive sampling algorithm. In the first step, noise is removed by the median-filtering technique in 2D image space. Since not all pixels can be used to reconstruct the 3D mesh due to limited system resources, the filtered map is adaptively sampled based on the depth variation. Sampled 2D pixels, called feature points, are triangulated and projected onto 3D space. In the second refinement step, we apply the Gaussian smoothing technique to the reconstructed 3D surface. Finally, 3D surfaces are rendered to compute a smooth depth map from Z-buffer.

In this paper, we propose a framework for virtual reality, I²-NEXT, which enables users to interact with virtual objects by tangible objects in immersive networked virtual environment. The primary goal of this framework is to support rapid development of immersive and interactive virtual reality systems as well as various types of user interfaces. The proposed framework consists of user interface for interactions, immersive virtual environment, and networking interface. In this framework, we adopt several design patterns to guarantee that either developers or users (artists) can easily implement their VR applications without strong knowledge of VR techniques such as programming, libraries etc. One of the key features of this framework is the presence of the device module which supports a natural user interaction in a virtual environment. For example, the proposed framework provides users with tangible objects so that the users are able to manipulate virtual objects by touching real objects. The proposed framework also supports large scale stereoscopic display through clustering technique. To realize the effectiveness of the proposed framework, we have been developing an application for digital heritage reconstruction. Having been through development of the system, we believe that virtual reality technology is one of the promising technologies which enable users to experience realities in a digital space. Detailed explanations of each component and system architecture are presented.

In this paper, we present the Tangible Media Control System (TMCS), which allows users to manipulate media contents through physical objects in an intuitive way. Currently, most people access digital media contents by exploiting GUI. However, it only provides limited manipulation of the contents. The proposed system, instead of a mouse and a keyboard, adopts two types of tangible objects, i.e. a RFID-enabled object and a tracker-embedded object. The TMCS enables users to easily access and control digital media contents through tangible objects. In addition, it supports an interactive media controller which can be used to synthesize media contents according to users' taste. It also offers personalized contents, which suits users' preferences, by exploiting context such as the users' profile and situational information. Accordingly, the TMCS demonstrates that tangible interfaces with context can provide more effective interfaces to satisfy users' demands. Therefore, the proposed system can be applied to various interactive applications such as multimedia education, entertainment, multimedia editor, etc.

In this paper, a novel projection-based method is presented to register partial 3D point clouds, acquired from a multi-view camera, for 3D reconstruction of an indoor scene. In general, conventional registration methods for partial 3D point clouds require a high computational complexity and much time for registration. Moreover, these methods are not robust for 3D point cloud which has a low precision. To overcome these drawbacks, a projection-based registration method is proposed. Firstly, depth images are refined based on both temporal and spatial properties. The former involves excluding 3D points with large variation, and the latter fills up holes referring to four neighboring 3D points, respectively. Secondly, 3D point clouds acquired from two views are projected onto the same image plane, and two-step integer mapping is applied to search for correspondences through the modified KLT. Then, fine registration is carried out by minimizing distance errors based on adaptive search range. Finally, we calculate a final color referring to the colors of corresponding points and reconstruct an indoor scene by applying the above procedure to consecutive scenes. The proposed method not only reduces computational complexity by searching for correspondences on a 2D image plane, but also enables effective registration even for 3D points which have a low precision. Furthermore, only a few color and depth images are needed to reconstruct an indoor scene. The generated model can be adopted for interaction with as well as navigation in a virtual environment.

Our group is exploring interactive multi- and hypermedia, especially applied to virtual and mixed reality multimodal groupware systems. We are researching user interfaces to control source→sink transmissions in synchronous groupware (like teleconferences, chatspaces, virtual concerts, etc.). We have developed two interfaces for privacy visualization of narrowcasting (selection) functions in collaborative virtual environments (CVES): for a workstation WIMP (windows/icon/menu/pointer) GUI (graphical user interface), and for networked mobile devices, 2.5- and 3^rd-generation mobile phones. The interfaces are integrated with other CVE clients, interoperating with a heterogeneous multimodal groupware suite, including stereographic panoramic browsers and spatial audio backends & speaker arrays. The narrowcasting operations comprise an idiom for selective attention, presence, and privacy-- an infrastructure for rich conferencing capability.

In this paper, we propose a novel virtual display system for a real object surface by using a video projector, so that the viewer can feel as if digital images are printed on the real surface with arbitrary shape. This system consists of an uncalibrated camera and video projector connected to a same PC and creates a virtual object by rendering 2D contents preserved beforehand onto a white object in a real world via a projector. For geometry registration between the rendered image and the object surface correctly, we regard the object surface as a set of a number of small rectangular regions and perform geometry registration by calculating homographies between the projector image plane and the each divided regions. By using such a homography-based method, we can avoid calibration of a camera and a projector that is necessary in a conventional method. In this system, we perform following two processes. First of all, we acquire the status of the object surface from images which capture the scene that color-coded checker patterns are projected on it and generate image rendered on it without distortion by calculating homographies. After once the projection image is generated, the rendered image can be updated if the object surface moves, or refined when it is stationary by observing the object surface. By this second process, the system always offers more accurate display. In implementation, we demonstrate our system in various conditions. This system enables it to project them as if it is printed on a real paper surface of a book. By using this system, we expect the realization of a virtual museum or other industrial application.

Most of the traveller guidance services (TGS) are based on GPS technology and generally concerned with the position data mapping on the simplified 2D electronic map in order to provide macro level service facility such as drive direction notifications. Digital GIS based GPS entails in situ intuitive visualization. The visually enhanced TGS can improve the global and local awareness of unknown areas. In this paper, we propose a reliable new TGS system that provides 3D street as well as pin-pointed destination information in two stages of its interactive services; web-based and AR-based. The web server generates a guiding path on 2D digital map and displays 3D car-driving animation along the path. And, the AR-based service is embedded so that users can interactively obtain the detailed micro-level information of a specific section in the area with their fingertips. The implementation is based on autoformation of on-line GIS data structures from the available priori. For the verification, a 54 road network is selected as a test area. In the service demonstration, we show the effective awareness of street environments and the usefulness of this new TGS system.

In this paper, we demonstrate an immersive and interactive broadcasting production system with a new haptically enhanced multimedia broadcasting chain. The system adapts Augmented Reality (AR) techniques, which merges captured videos and virtual 3D media seamlessly through multimedia streaming technology, and haptic interaction technology in near real-time. In this system, viewers at the haptic multimedia client can interact with AR broadcasting production transmitted via communication network. We demonstrate two test applications, which show that the addition of AR- and haptic-interaction to the conventional audio-visual contents can improve immersiveness and interactivity of viewers with rich contents service.

This paper provides an overview of identity management with emphasis on the federated approaches, covering related standard specifications, and security and privacy considerations. Actual business cases, ranging from B2E to B2C in different industries, such as mobile communications, gaming, and digital TV are introduced. The Liberty Alliance identity frameworks are used as a concrete example to illustrate federated identity management technologies. Moreover, other approaches, such as SAML, shibboleth and WS-Federation, are compared.

We have proposed Rate-based Multi-path Transmission Control Protocol (R-M/TCP) for improving reliability and performance of data transfer over the Internet by using multiple paths. Congestion control in R-M/TCP is performed in a rate-based and loss-avoidance manner. It attempts to estimate the available bandwidth and the queue length of the used routes in order to fully utilize the bandwidth resources. However, it has been reported that when the used routes' characteristics, i.e. available bandwidth and delay, are much different, R-M/TCP cannot achieve the desired throughput from the routes. This is because R-M/TCP originally transmits data packets in a round-robin manner through the routes. In this paper, therefore, we propose R-M/TCP using Packet Scheduling Algorithm (PSA). Instead of using the round-robin manner, R-M/TCP utilizes PSA that accounts for time-varying bandwidth and delay of each path so that number of data packets arriving in out-of-order at the receiver can be minimized and the desired throughput can be achieved. Quantitative simulations are conducted to show effectiveness of R-M/TCP using PSA.

The queue size at a bottleneck would impact the performance of TCP protocols, especially when running a single TCP flow in networks with a large bandwidth-delay product. However, queue size has been not well considered in experiments. This paper shows how bottleneck queue size influences TCP protocols performance. Bursityness of advanced TCPs is examined. Ways of estimating queue size are introduced. Sending a UDP packet train until a loss is detected is a method to measure queuing delay to estimate queue size. Watching a loss in a TCP session to measure round trip time and calculate the queue size is also discussed. Results from experiments with a network emulator and a real network are reported. The results indicated that a layer-2 switch at a congestion point would be a major factor of decreasing TCP performance in a fast long distant path.

This paper presents a novel analytical approach to evaluate the signaling load of Mobile IPv6 (MIPv6) and Hierarchical Mobile IPv6 (HMIPv6). Previous analytical approaches for IP mobility management have not provided a complete and general framework for the performance analysis; no consideration of either periodic binding refresh cost or extra packet tunneling cost from the viewpoint of IP mobility management, and no in-depth investigation with respect to various system parameters. In this paper, according to the proposed analytical approach, we derive the location update costs (i.e., the sum of binding update costs and binding refresh costs), packet tunneling costs, inside-domain signaling costs, outside-domain signaling costs, and total signaling costs, which are generated by a mobile node (MN) during its average domain residence time in case MIPv6 or HMIPv6 is deployed under the same network architecture, respectively. Moreover, based on these derived costs, we evaluate the impacts of various system parameters on the signaling costs generated by an MN in MIPv6 and HMIPv6. The aim of this paper is not to determine which protocol performs better, but evaluate the performance that can be expected for each protocol under the various conditions, broaden our deep understanding of the various parameters that may influence the performance, and provide insight for the deployment of the two protocols.

Mobile devices have acceleratedly penetrated into our daily lives. Though they were originally designed as a communication tool or for personal use, and due to the rapid availability of wireless network technologies, people have begun to use mobile devices for supporting collaborative work and learning. There is, however, a serious problem in mobile devices related to their user interfaces. In this paper, we try to alleviate the problem and propose intuitive techniques for information transfer, which is one of the typical usages of mutually-connected computers. Our system, Toss-It, enables a user to send information from the user's PDA to other electronic devices with a "toss" or "swing" action, like a user would toss a ball or deal cards to others. The implementation of Toss-It consists of three principle parts - gesture recognition, location recognition, and file transfer. We especially describe the details of gesture recognition and location recognition. We then evaluate the practicability and usability of Toss-It through the experiments. We also discuss user scenarios describing how Toss-It can support users' collaborative activities.

The IETF (Internet Engineering Task Force) has developed a Network Mobility (NEMO) basic support protocol by extending the operation of Mobile IPv6 to provide uninterrupted Internet connectivity to the communicating nodes of mobile networks. The protocol uses a mobile router (MR) in the mobile network to perform prefix scope binding updates with its home agent (HA) to establish a bi-directional tunnel between the HA and MR. This solution reduces location-update signaling by making network movements transparent to the mobile nodes behind the MR. However, delays in data delivery and higher overheads are likely to occur because of sub-optimal routing and multiple encapsulation of data packets. To resolve these problems, we propose a mobile router-assisted route optimization (MoRaRo) scheme for NEMO support. With MoRaRo, a mobile node performs route optimization with a correspondent node only once, at the beginning of a session. After that the MR performs route optimization on behalf of all active mobile nodes when the network moves. The virtue of this scheme is that it requires only slight modification of the implementation of the NEMO basic support protocol at local entities such as the MR and mobile nodes of the mobile network, leaving entities in the core or in other administrative domains untouched. MoRaRo enables a correspondent node to forward packets directly to the mobile network without any tunneling, thus reducing packet delay and encapsulation overheads in the core network. To enable the scheme to be evaluated, we present the results of both theoretical analysis and simulation.

In a Peer-to-Peer (P2P) network, in order to improve the search performance and to achieve load balancing, replicas of original data are created and distributed over the Internet. However, the replication methods which have been proposed so far focus only on the improvement of search performance. In this paper, we examine the load on the storage systems, which is due to writing and reading, and propose two replication methods for balancing the load on the storages distributed over P2P networks while limiting the degradation of the search performance within an acceptable level. Furthermore, we investigate the performance of our proposed replication methods through computer simulations, and show their effectiveness in balancing the load.

A peer-to-peer complementary currency can be a powerful tool for promoting exchanges and building relationships on the Internet. iWAT [1] is a proposed such currency based on the WAT System [2], a polycentric complementary currency using WAT tickets as its media of exchange: participants spontaneously issue and circulate the tickets as needed, whose values are backed up by chains of trust. iWAT implements the tickets electronically by exchanging messages signed in OpenPGP [3]. This paper investigates an extension to the design of iWAT to facilitate mutual help among peers in need. In particular, we investigate additional "reduction" tickets whose values are reduced over time. By deferring redemption of such tickets, the participants can contribute to reduce the debts of the issuers, and the issuers help participants by providing exchange media that accelerate spending. This paper describes in detail how incentive-compatibility is achieved by this extended design; we predict that the following properties will hold, which resulted from a game-theoretical analysis.
1. Rapid circulation, or a reduction ticket will typically circulate at high speed until its effective value reaches the scheduled minimum, and
2. Vanishment equilibrium, or the system will be most stable if the values of tickets are to be reduced down to zero.
A reference implementation of iWAT has been developed in the form of a plug-in for an XMPP [4],[5] instant messaging client. We have been putting the currency system into practical use, to which the proposed feature has been added.

Security protocols flaws represent a substantial portion of security exposures of data networks. In order to evaluate security protocols against any attack, formal methods are equipped with a number of techniques. Unfortunately, formal methods are applicable for static state only, and don't guarantee detecting all possible flaws. Therefore, formal methods should be complemented with dynamic protection. Anomaly detection systems are very suitable for security protocols environments as dynamic activities protectors. This paper presents an intrusion detection system that uses a number of different anomaly detection techniques to detect attacks against security protocols.

Machine learning and data mining algorithms are increasingly being used in the intrusion detection systems (IDS), but their performances are laggard to some extent especially applied in network based intrusion detection: the larger load of network traffic monitoring requires more efficient algorithm in practice. In this paper, we propose and design an anomaly intrusion detection (AID) system based on the vector quantization (VQ) which is widely used for data compression and high-dimension multimedia data index. The design procedure optimizes the performance of intrusion detection by jointly accounting for accurate usage profile modeling by the VQ codebook and fast similarity measures between feature vectors to reduce the computational cost. The former is just the key of getting high detection rate and the later is the footstone of guaranteeing efficiency and real-time style of intrusion detection. Experiment comparisons to other related researches show that the performance of intrusion detection is improved greatly.

Contents delivery services are deployed over the Internet and its technology. Contents delivery services demand high quality. Consequently, a large network capacity is required. In order to efficiently deliver such contents and to meet high quality demands, network operators are required to explicitly indicate the branch node and/or the link so that packets are efficiently sent. Fast restoration at the time of trouble has become an important issue. MPLS technology has been utilized to realize Traffic Engineering and Fast Reroute to cope with this issue. However, only a point-to-point path is allowed in the current MPLS technology. To allow an efficient high-quality contents delivery, a point-to-multipoint path called as P2MP-LSP is under discussion. However, proposed methods lack of getting information about P2MP-LSP node due to the current MPLS signaling method to establish P2MP-LSP. This paper introduces a discovery technique of the P2MP-LSP node and an establishment technique of the P2MP-LSP using BGP. In addition, a basic function of this proposal is experimented, and its practicality is evaluated.

To achieve profits derived from multihoming, policy routing is supposed to be necessary. For routers' policy routing, some systems and implementations already exist. In addition to routers' operating systems with policy routing functions, IP Filter and iproute2 are useful for policy routing on routers. But they don't suit for policy routing on multihomed end-hosts because of differences on network environments between routers and end-hosts. In contrast with routers' network environments, there are some dynamic changes of network environments with some types of end-hosts, for example, laptop computers. Therefore, another policy routing system is needed for end-hosts' policy routing, which adapts to dynamic changes of network environment. PMPATH (Policy based MultiPATH routing system) is a policy routing system and designed especially for end-hosts uses. PMPATH uses source address selection as network selection, PMPATH can adapt to dynamic changes of network environment. PMPATH's policies for source address selection are defined in form of address block. PMPATH also provides outgoing interface selection. We implemented PMPATH on NetBSD 2.99.15 and tested its performance on multihomed end-hosts. PMPATH implementation works well even if there are changes of network environment. In this result, it is shown that PMPATH is useful for multihomed end-hosts' policy routing.

With great advances in the mobile technology and wireless communications, users expect to be online anytime anywhere. However, due to the high cost of being online, applications are still implemented as partially connected to the server. In many data-intensive mobile client/server frameworks, it is a daunting task to archive and index such a mass volume of complex data that are continuously added to the server when each mobile client gets online. In this paper, we propose a scalable technique called Seeded Clustering that allows us to maintain R-tree indexes by bulk insertion while keeping pace with high data arrival rates. Our approach uses a seed tree, which is copied from the top k levels of a target R-tree, to classify input data objects into clusters. We then build an R-tree for each of the clusters and insert the input R-trees into the target R-tree in bulk one at a time. We present detailed algorithms for the seeded clustering and bulk insertion as well as the results from our extensive experimental study. The experimental results show that the bulk insertion by seeded clustering outperforms the previously known methods in terms of insertion cost and the quality of target R-trees measured by their query performance.

XML query processing is one of the most active areas of database research. Although the main focus of past research has been the processing of structural XML queries, there are growing demands for a full-text search for XML documents. In this paper, we propose XICS (XML Indices for Content and Structural search), which aims at high-speed processing of both full-text and structural queries in XML documents. An important design principle of our indices is the use of a B⁺-tree. To represent the structural information of XML trees, each node in the XML tree is labeled with an identifier. The identifier contains an integer number representing the path information from the root node. XICS consist of two types of indices, the COB-tree (COntent B⁺-tree) and the STB-tree (STructure B⁺-tree). The search keys of the COB-tree are a pair of text fragments in the XML document and the identifiers of the leaf nodes that contain the text, whereas the search keys of the STB-tree are the node identifiers. By using a node identifier in the search keys, we can retrieve only the entries that match the path information in the query. The STB-tree can filter nodes using structural conditions in queries, while the COB-tree can filter nodes using text conditions. We have implemented a COB-tree and an STB-tree using GiST and examined index size and query processing time. Our experimental results show the efficiency of XICS in query processing.

Management of applications in the new world of pervasive computing requires new mechanisms to be developed for admission control, QoS negotiation, allocation and scheduling. To solve such resource-allocation and QoS provisioning problems within pervasive and ubiquitous computational environments, distribution and decomposition of the computation are important. In this paper we present a QoS-based welfare economic resource management model that models the actual price-formation process of an economy. We compare our economy-based approach with a mathematical approach we previously proposed. We use the constructs of application benefit functions and resource demand functions to represent the system configuration and to solve the resource allocation problems. Finally empirical studies are conducted to evaluate the performance of our proposed pricing model and to compare it with other approaches such as priority-based scheme and greedy method.

Digital watermarks on pictures are more useful when they are better able to survive image processing operations and when they cause less degradation of picture quality. Random geometric distortion is one of the most difficult kinds of image processing for watermarks to survive because of the difficulty of synchronizing the expected watermark patterns to the watermarks embedded in pictures. This paper proposes three methods to improve a previous method that is not affected by this difficulty but that is insufficient in maintaining picture quality and treating other problems in surviving image processing. The first method determines the watermark strength in L*u*v* space, where human-perceived degradation of picture quality can be measured in terms of Euclidian distance, but embeds and detects watermarks in YUV space, where the detection is more reliable. The second method, based on the knowledge of image quantization, uses the messiness of color planes to hide watermarks. The third method reduces detection noises by preprocessing the watermarked image with orientation-sensitive image filtering, which is especially effective in picture portions where pixel values change drastically. Subjective evaluations have shown that these methods improved the picture quality of the previous method by 0.5 point of the mean evaluation score at the representative example case. On the other hand, the watermark strength of the previous method could be increased by 30% through 60% while keeping the same picture quality. Robustness to image processing has been evaluated for random geometric distortion, JPEG compression, Gaussian noise addition, and median filtering and it was clarified that these methods reduced the detection error ratio to 1/10 through 1/4. These methods can be applied not only to the previous method but also to other types of pixel-domain watermarking such as the Patchwork watermarking method and, with modification, to frequency-domain watermarking.

To support international medical collaboration activities carried out among Japan, Sri Lanka and other countries in Asia, we have been developing a collaboration system, which incorporates synchronous and asynchronous collaboration paradigms through network-effective persistent information sharing. The designed system facilitates synchronous collaborative work on maxillo-facial cases through real-time high-quality image delivery and by bringing the system database objects to a shared-workspace. The asynchronous activities are supported through a web based collaborative environment that enables both activity-centric and object-centric navigation of collaboration contents incorporated with email notifications of the system updates. Multimedia records of synchronous sessions are posted to the system database for later reviewing. Geographically dispersed groups experiencing different scales of network heterogeneities are served by a distributed application level multicast overlay and an adaptive multimedia delivery mechanism. The designed collaboration system also features several useful collaboration tools, user friendly GUI and ubiquitous connectivity.

This paper proposes an efficient technique for human motion recognition based on motion history images and an eigenspace technique. In recent years, human motion recognition has become one of the most popular research fields. It is expected to be applied in a security system, man-machine communication, and so on. In the proposed technique, we use two feature images and the eigenspace technique to realize high-speed recognition. An experiment was performed on recognizing six human motions and the results showed satisfactory performance of the technique.

To find the best transformation between a "model" point set and an "image" point set is the main purpose of point pattern matching. The similarity measure plays a pivotal role and is used to determine the degree of resemblance between two objects. Although some well-known Hausdorff distance measures work well for this task, they are very computationally expensive and suffer from the noise points. In this paper, we propose a novel similarity measure using the Hilbert curve named Hilbert scanning distance (HSD) to resolve the problems. This method computes the distance measure in the one-dimensional (1-D) sequence instead of in the two-dimensional (2-D) space, which greatly reduces the computational complexity. By applying a threshold elimination function, large distance values caused by noise and position errors (e.g. those that occur with feature or edge extraction) are removed. The proposed algorithm has been applied to the task of matching edge maps with noise. The experimental results show that HSD can provide sufficient information for image matching within low computational complexity. We believe this sets a new direction for the research of point pattern recognition.

Syntax and prosody are closely related to each other. This paper is concerned with the problem of exploiting pause information for recovering dependency structures of read Japanese sentences. Our parser can handle both symbolic information such as dependency rule and numerical information such as the probability of dependency distance of a phrase in a unified way as linguistic information. In our past work, post-phrase pause that immediately succeeds a phrase in question was employed as prosodic information. In this paper, we employed two kinds of pauses in addition to the post-phrase pause: post-post-phrase pause that immediately succeeds the phrase that follows a phrase in question, and pre-phrase pause that immediately precedes a phrase in question. By combining the three kinds of pause information linearly with the optimal combination weights that were determined experimentally, the parsing accuracy was improved compared to the case where only the post-phrase pause was used as in our previous work. Linear combination of pause and fundamental frequency information yielded further improvement of parsing accuracy.

In this paper, a revised method is proposed in order to determine the parameters of an F₀ generation model from the observed F₀ contour automatically. Compared with the previous method, there are two points revised in the proposed method. Firstly, we relax the endpoint constraint in the dynamic programming method, especially we allow the timing of the first phrase command to be earlier than the beginning point of the actual F₀ pattern. Secondly, the z-transform method is introduced to convert the equation of the F₀ model in order to simplify the calculation and save the computation time. An experiment with 100 sentences spoken by two males and two females selected from the speech database "ATR 503 sentences" has shown that the proposed method is effective as we expected.

This paper presents an approach that uses the Viterbi algorithm in a stereo correspondence problem. We propose a matching process which is visualized as a trellis diagram to find the maximum a posterior result. The matching process is divided into two parts: matching the left scene to the right scene and matching the right scene to the left scene. The last result of stereo problem is selected based on the minimum error for uniqueness by a comparison between the results of the two parts of matching process. This makes the stereo matching possible without explicitly detecting occlusions. Moreover, this stereo matching algorithm can improve the accuracy of the disparity image, and it has an acceptable running time for practical applications since it uses a trellis diagram iteratively and bi-directionally. The complexity of our proposed method is shown approximately as O(N²P), in which N is the number of disparity, and P is the length of the epipolar line in both the left and right images. Our proposed method has been proved to be robust when applied to well-known samples of stereo images such as random dot, Pentagon, Tsukuba image, etc. It provides a 95.7 percent of accuracy in radius 1 (differing by 1) for the Tsukuba images.

In this paper, we propose a construction method of three-dimensional deformable models that represent tree-shaped human organs, such as bronchial tubes, based on results obtained by statistically analyzing the distributions of bifurcation points in the tree-shaped organs. The models are made to be used as standard templates of tree-shaped organs in medical image recognition, and are formed by control points that can be uniquely identified as structural elements of organs such as bifurcation tracheae in bronchial tubes. They can be transfigured based on the statistical validity of relationships between the control points. The optimal state of that transfiguration is determined within the framework of energy minimization. Experimental results from bronchial tubes are shown on actual CT images.

Phonetic string search of written text is an important topic in Information Retrieval. While most of the previous methods convert a string into intermediate codes with phonetic transformation rules, this paper proposes a novel algorithm to segment two phonetic strings into syllables and find the optimal pairing of the corresponding syllables to calculate their similarity score. The experiment shows that this method is very effective and flexible. It can be easily adapted to different datasets and achieves optimal performance on average.

Finding intracranial aneurysms plays a key role in preventing serious cerebral diseases such as subarachnoid hemorrhage. For detection of aneurysms, magnetic resonance angiography (MRA) can provide detailed images of arteries non-invasively. However, because over 100 MRA images per subject are required to cover the entire cerebrum, image diagnosis using MRA is very time-consuming and labor-intensive. This article presents a computer-aided diagnosis (CAD) system for finding aneurysms with MRA images. The principal components are identification of aneurysm candidates (= ROIs; regions of interest) from MRA images and estimation of a fuzzy degree for each aneurysm candidate based on a case-based reasoning (CBR). The fuzzy degree indicates whether a candidate is true aneurysm. Our system presents users with a limited number of ROIs that have been sorted in order of fuzzy degree. Thus, this system can decrease the time and the labor required for detecting aneurysms. Experimental results using phantoms indicate that the system can detect all aneurysms at branches of arteries and all saccular aneurysms produced by dilation of a straight artery in 1 direction perpendicular to the principal axis. In a clinical evaluation, performance in finding aneurysms and estimating the fuzzy degree was examined by applying the system to 16 subjects with a total of 19 aneurysms. The experimental results indicate that this CAD system detected all aneurysms except a fusiform aneurysm, and gave high fuzzy degrees and high priorities for the detected aneurysms.

IETF MANET (Mobile Ad Hoc Network) WG has standardized OLSR (Optimized Link State Routing) as its proactive routing protocol. In addition, TCP (Transmission Control Protocol) is still needed for MANET thanks to its suitability for smooth integration with the fixed Internet. In particular, TCP-Vegas is a well-known transport protocol that can efficiently take account of network condition. However, TCP-Vegas that requires an accurate BaseRTT estimation cannot be directly applied to MANET because a route change makes the estimated BaseRTT obsolete. We propose a technique to improve the performance of TCP-Vegas by considering the route change, and show the performance improvement through simulation study using the ns-2 simulator.

In this paper, we propose a new clustered reconfigurable interconnect network (CRIN) BIST that can improve the embedding probabilities of random-pattern-resistant-patterns. A simulated annealing based algorithm that maximizes the embedding probabilities of scan test cubes has been developed to reorder scan cells. Experimental results demonstrate that the proposed CRIN BIST technique reduces test time by 35% and the storage requirement by 39% in comparison with previous work.

An adaptive data hiding scheme capable of hiding considerable quantities of secret data while preserving acceptable visual quality for cover images is proposed. The major idea of this scheme is to hide secret data into the compressed codes of cover image during the encoding process of side-match vector quantization (SMVQ) such that the interceptors will not capture the secret information. Based on the experimental results, it is confirmed that the proposed scheme is better than earlier works. Moreover, the receiver can efficiently receive both the compressed cover image and the hidden secret data at the same time.

An efficient and simple approach to consonant/vowel (C/V) segmentation by incorporating the SNR improvement of a speech enhancement system with the energy variation of two adjacent frames is proposed. Experimental results show that the proposed scheme performs well in segmenting C/V for a spontaneously spoken utterance.

The deblocking filter in H.264 is an efficient tool to reduce blocking artifact, but it also blurs the details or retains blocking artifact perceptible in some high-activity areas. In this paper, we improve the filtered pixel classification and filtering schemes used by the deblocking filter in H.264 to keep the sharpeness of real edges and minimize over-smoothing.

In this paper, we present the biometric authentication system based on the fusion of two user-friendly biometric modalities: Iris and Face. Using one biometric feature can lead to good results, but there is no reliable way to verify the classification. To achieve robust identification and verification we are combining two different biometric features. We specifically apply 2-D discrete wavelet transform to extract the feature sets of low dimensionality from the iris and face. And then to obtain Reduced Joint Feature Vector (RJFV) from these feature sets, Direct Linear Discriminant Analysis (DLDA) is used in our multimodal system. This system can operate in two modes: to identify a particular person or to verify a person's claimed identity. Our results for both cases show that the proposed method leads to a reliable person authentication system.

Since sentences are the basic propositional units of text, knowing their themes should help in completing various tasks such as automatic summarization requiring the knowledge about the semantic content of text. Despite the importance of determining the theme of a sentence, however, few studies have investigated the problem of automatically assigning a theme to a sentence. In this paper, we examine the notion of sentence theme and propose an automatic scheme where head-driven patterns are used for theme assignment. We tested our scheme with sentences in encyclopedia articles and obtained a promising result of 98.96% in F-score for training data and 88.57% for testing data, which outperform the baseline using all but the head-driven patterns.

An efficient clustering strategy for estimation of distribution algorithms (EDAs) is presented. It is used for properly fitting probabilistic models that play an important role in guiding search direction. To this end, a fitness-aided ordering scheme is devised for deciding the input sequence of samples (i.e., individuals) for clustering. It can effectively categorise the individuals by using the (available) information about fitness landscape. Moreover, a virtual leader is introduced for providing a reliable reference for measuring the distance from samples to its own cluster. The proposed algorithm incorporates them within the framework of random the leader algorithm (RLA). Experimental results demonstrate that the proposed approach is more effective than the existing ones with regard to probabilistic model fitting.