We consider the problem of finding the minimizer of a function f:ℝd→ℝ of the finite-sum form minf(w)=1/n∑nifi(w). This problem has been studied intensively in recent years in the field of machine learning (ML). One promising approach for large-scale data is to use a stochastic optimization algorithm to solve the problem. SGDLibrary is a readable, flexible and extensible pure-MATLAB library of a collection of stochastic optimization algorithms. The purpose of the library is to provide researchers and implementers a comprehensive evaluation environment for the use of these algorithms on various ML problems.

Stochastic variance reduction algorithms have recently become popular for minimizing the average of a large, but finite number of loss functions. The present paper proposes a Riemannian stochastic quasi-Newton algorithm with variance reduction (R-SQN-VR). The key challenges of averaging, adding, and subtracting multiple gradients are addressed with notions of retraction and vector transport. We present convergence analyses of R-SQN-VR on both non-convex and retraction-convex functions under retraction and vector transport operators. The proposed algorithm is evaluated on the Karcher mean computation on the symmetric positive-definite manifold and the low-rank matrix completion on the Grassmann manifold. In all cases, the proposed algorithm outperforms the state-of-the-art Riemannian batch and stochastic gradient algorithms.

Stochastic variance reduction algorithms have recently become popular for minimizing the average of a large, but finite number of loss functions on a Riemannian manifold. The present paper proposes a Riemannian stochastic recursive gradient algorithm (R-SRG), which does not require the inverse of retraction between two distant iterates on the manifold. Convergence analyses of R-SRG are performed on both retractionconvex and non-convex functions under computationally efficient retraction and vector transport operations. The key challenge is analysis of the influence of vector transport along the retraction curve. Numerical evaluations reveal that R-SRG competes well with state-of-the-art Riemannian batch and stochastic gradient algorithms.

Sequential data modeling and analysis have become indispensable tools for analyzing sequential data, such as time-series data, because larger amounts of sensed event data have become available. These methods capture the sequential structure of data of interest, such as input-output relations and correlation among datasets. However, because most studies in this area are specialized or limited to their respective applications, rigorous requirement analysis of such models has not been undertaken from a general perspective. Therefore, we particularly examine the structure of sequential data, and extract the necessity of "state duration" and "state interval" of events for efficient and rich representation of sequential data. Specifically addressing the hidden semi-Markov model (HSMM) that represents such state duration inside a model, we attempt to add representational capability of a state interval of events onto HSMM. To this end, we propose two extended models: an interval state hidden semi-Markov model (IS-HSMM) to express the length of a state interval with a special state node designated as "interval state node"; and an interval length probability hidden semi-Markov model (ILP-HSMM) which represents the length of the state interval with a new probabilistic parameter "interval length probability." Exhaustive simulations have revealed superior performance of the proposed models in comparison with HSMM. These proposed models are the first reported extensions of HMM to support state interval representation as well as state duration representation.

Sequential data modeling and analysis have become indispensable tools for analyzing sequential data, such as time-series data, because larger amounts of sensed event data have become available. These methods capture the sequential structure of data of interest, such as input-output relations and correlation among datasets. However, because most studies in this area are specialized or limited to their respective applications, rigorous requirement analysis of such models has not been undertaken from a general perspective. Therefore, we particularly examine the structure of sequential data, and extract the necessity of “state duration” and “state interval” of events for efficient and rich representation of sequential data. Specifically addressing the hidden semi-Markov model (HSMM) that represents such state duration inside a model, we attempt to add representational capability of a state interval of events onto HSMM. To this end, we propose two extended models: an interval state hidden semi-Markov model (IS-HSMM) to express the length of a state interval with a special state node designated as “interval state node”
and an interval length probability hidden semi-Markov model (ILP-HSMM) which represents the length of the state interval with a new probabilistic parameter “interval length probability.” Exhaustive simulations have revealed superior performance of the proposed models in comparison with HSMM. These proposed models are the first reported extensions of HMM to support state interval representation as well as state duration representation.

This paper addresses network anomography, that is, the problem of inferring network-level anomalies from indirect link measurements. This problem is cast as a low-rank subspace tracking problem for normal flows under incomplete observations and an outlier detection problem for abnormal flows. Since traffic data is large-scale time-structured data accompanied with noise and outliers under partial observations, an efficient modeling method is essential. To this end, this paper proposes an online subspace tracking of a Hankelized time-structured traffic tensor for normal flows based on the Candecomp/PARAFAC decomposition exploiting the recursive least squares algorithm. We estimate abnormal flows as outlier sparse flows via sparsity maximization in the underlying under-constrained linear-inverse problem. A major advantage is that our algorithm estimates normal flows by low-dimensional matrices with time-directional features as well as the spatial correlation of multiple links without using the past observed measurements and the past model parameters. Extensive numerical evaluations show that the proposed algorithm achieves faster convergence per iteration of model approximation and better volume anomaly detection performance compared to state-of-the-art algorithms.

We propose a novel Riemannian manifold preconditioning approach for the tensor completion problem with rank constraint. A novel Riemannian metric or inner product is proposed that exploits the least-squares structure of the cost function and takes into account the structured symmetry that exists in Tucker decomposition. The specific metric allows to use the versatile framework of Riemannian optimization on quotient manifolds to develop preconditioned nonlinear conjugate gradient and stochastic gradient descent algorithms in batch and online setups, respectively. Concrete matrix representations of various optimization-related ingredients are listed. Numerical comparisons suggest that our proposed algorithms robustly outperform state-of-the-art algorithms across different synthetic and real-world datasets.

Analysis of sequential event data has been recognized as one of the essential tools in data modeling and analysis field. In this paper, after the examination of its technical requirements and issues to model complex but practical situation, we propose a new sequential data model, dubbed Duration and Interval Hidden Markov Model (DI-HMM), that efficiently represents 'state duration' and 'state interval' of data events. This has significant implications to play an important role in representing practical time-series sequential data. This eventually provides an efficient and flexible sequential data retrieval. Numerical experiments on synthetic and real data demonstrate the efficiency and accuracy of the proposed DI-HMM.

The background, the proposed concept and design, and the result of the research and development work on the "relational-metric based new generation network" will be presented. The relational metric technique enables numerically modeling social distance between people, things and places from various types of life-log and sensor data. The integration of this and network virtualization enables us to control 'social distance of information for people' as people desire, which has been validated from experiment results.

Analysis of sequential event data has been recognized as one of the essential tools in data modeling and analysis field. In this paper, after the examination of its technical requirements and issues to model complex but practical situation, we propose a new sequential data model, dubbed Duration and Interval Hidden Markov Model (DI-HMM), that efficiently represents "state duration" and "state interval" of data events. This has significant implications to play an important role in representing practical time-series sequential data. This eventually provides an efficient and flexible sequential data retrieval. Numerical experiments on synthetic and real data demonstrate the efficiency and accuracy of the proposed DI-HMM.

This chapter tackles two problems we face when extracting values from sensing data: 1) it is hard for humans to understand raw/unprocessed sensing data and 2) it is inefficient in terms of management costs to keep all sensing data 'usable'. This chapter also discusses a solution, i.e., the socialized system, which encodes the characteristics of sensing data in relational graphs so as to extract values that originally contained the sensing data from the relational graphs. The system model, the encoding/decoding logic, and the real-dataset examples are presented. We also propose a content distribution paradigm built on the socialized system that is called SocialCast. SocialCast can achieve load balancing, low-retrieval latency, and privacy while distributing content using relational metrics produced from the relational graph of the socialized system. We did a simulation and present the results to demonstrate the effectiveness of this approach.

There is a fact that, when people send messages, the destinations are not always deterministic; people sometimes want to send a message to someone who are just interested in the content of the message. However, e-mail services, which have been recognized as the de-facto standard for years, do not have such capability though a couple of emerging messaging applications like micro-blogging have it. Problems happen particularly when people use the e-mail system to make an announcement regarding an official event, which would be easily received even by people who are not likely interested in it. This paper proposes a new e-mail system that has the capability of dynamically controlling recipients of messages according to the context.

This paper presents a proposal of a music image generator for users to grasp music easily and instantaneously on consumer electronics devices such as music players and smartphones. It provides users faster and more efficient music browsing functionality using a technique that automatically generates a music image that matches lyric contents. More specifically, a scene theme such as an event or place is extracted from lyrics using text mining techniques. A suitable music image is then created dynamically to match this extracted scene theme. This image generation is achieved by synthesizing multiple image materials to express the extracted scene theme flexibly. For choosing image materials, scene knowledge that has representative scene object compositions is proposed. Using this knowledge, the image synthesizer can select appropriate materials. Finally, the performance of the proposed automatic knowledge creation method is evaluated by comparison with manually created knowledge. Based on optimal parameters calculated in the evaluation, some examples of music images for well-known music lyrics are presented. In addition, the implemented system applying the proposed generator onto a server-side and a consumer-side consumer electronics device is also described.(1)

Recent wireless LAN technologies and service management technologies have produced widely diverse network services with no configuration task or effort for end-users' consumer electronics (CE) devices. This environment needs a mechanism that enables end-users to access services seamlessly across separated networks. To provide such accessibility, information for services in multiple networks should be available in each terminal that is currently connected to a wireless LAN. This paper presents a proposal of a new mechanism to collect and manage service information belonging to different WLANs autonomously for end-users. Specifically, an autonomous network switch and event-driven service information sharing scheme are proposed. The functionalities which are embedded on end-user CE devices require no central control device on a WLAN or the internet and require no modification of WLAN access points. The effectiveness of the proposal is demonstrated using simulation experiments. Future work is described(1).

High functionality and interactivity of video services are strongly desired. Earlier reports have presented an encoding scheme and a fast joining scheme for interactive multivision video streaming systems. Together, they enable numerous users to access any view area interactively at any desired resolution. The joining means the combination of multiple streams into a single video stream. The server joins pre-encoded multiple streams into one bitstream according to a user request; then the server transmits it simultaneously to many clients. This structure obviates complexity at the client terminal because the client terminal simply receives and decodes the single stream. This paper describes implementation of an actual HTTP-based streaming server by embedding the stream joiner module into a widely available HTTP server. The contribution of this paper is that the joining speed of the server surpasses the necessary speed attained with practical network equipment connecting to the server. Results of experiments described in this paper underscore the proposed method's effectiveness(1).

A new area-based distributed mobile cache system, a sustainable distributed Geocast technology, enables data caching temporarily in a designated local area using short-range communication technologies such as WiFi and Bluetooth. This system needs no infrastructure network and servers. This caching capability and mechanisms are expected to enhance short-range communications as a time-offset communication in a physical world. For that reason, it is sure to accelerate new location-based communication services. This paper reports a developed middleware for a widespread embedded operating system, and a software development kit with an open application programming interface. This middleware is expected to spur third-party developers to produce widely various applications using this proposed platform. Some pseudo-code is provided. Procedures to use this platform are also explained using an implemented geo-location social game(1).

This paper presents a proposal of a scheme to collect and manage content and service data belonging to separate multiple WLANs so that Mobile CE devices can use them. Content and service list sharing for seamless multiple WLANs access by network switching is proposed. This paper shows the effectiveness of the proposed algorithm using simulations, and describes future work(1).

This paper proposes music classification considering melody transition for browsing system to recommend tunes depending on users' preference. Especially, we apply the Prime form and the Interval-Class Vector to Melodies Markov Model proposed in [1](1).

High functionality and interactivity of video services are strongly desired. The authors have presented a proposal of an encoding scheme and a fast joining scheme for interactive multivision systems, where joining is the combination of multiple streams into a single video stream. This paper describes implementation of an actual HTTP-based streaming server by embedding the stream joiner module into a widely available HTTP server. The contribution of this paper is that the joining speed of the server surpasses the necessary speed attained with practical network equipment connecting to the server.

Our new distributed mobile cache system, a sustainable distributed Geocast technology, enables data caching temporarily in a designated local area. We released the open software development kit (SDK) for embedded systems. This paper explains details of the developed open SDK, an implementation guide, and three applications using this SDK1. © 2013 IEEE.

Multi-vision systems and panoramic video systems are expected to produce a new multimedia paradigm. Our previously presented multi-vision system allows many users to view multiple videos by virtue of fast stream joining algorithms. However image degradation occurs because these algorithms need restricted constant coding parameters. As described herein, we propose a QP control scheme to decrease this degradation without sacrificing high speed in the joiner. © 2013 IEEE.

This paper presents a proposal of a more resilient mechanism of an area-based sustainable cache system under a temporary decrease of the number of terminals. This mechanism suppresses interruptions of data relay and disappearances of cache data in the cache continually. The core technology enables each terminal to retain the received data in cache areas as "deactivated data" instead of deleting the data, even after a terminal leaves the area. Deactivated data are not relayed outside the target designated cache area. The data are reactivated and redistributed to others as relay data once the terminal revisits there. Consequently, the data are shared across time. Furthermore, particularly addressing the limitation of terminal storage, we proposed an adaptive scheme to select which data should be retained as deactivation data. This new mechanism leverages the proposed method under the storage limitation. As presented at the end of this paper, the simulation evaluations underscore the effectiveness of the proposed mechanism, and the adaptive selective scheme of deactivated data.

The concept of context awareness is believed to be a key enabler for the new ubiquitous network service paradigm brought by cloud computing platforms and smartphone OSs. In particular, autonomous context-based service customization is becoming an essential tool in this context because users cannot be expected to pick step by step the appropriate network services by manually and explicitly matching preferences for their current context. In this work, we hence focus on the core problem of how to detect changes of context for network services. In turn, detection of such changes can trigger timely system reconfigurations. We introduce a trigger detection mechanism based on a mixed graph-based representation model able to encode geographical and social relationships among people and social objects like stores, restaurants, and event spots. Our mechanism generates a trigger when a significant change in the graph takes place, and it is able to render significant changes in a geographical relationship that holds among objects socially connected with each other. The main benefits of our method are that (1) it does not require building reference models in advance, and (2) it can deal with different kinds of social objects uniformly once the graph is defined. A computer simulation scenario provides evidence on the expected performance of our method.

Multi-vision systems and panoramic video systems are expected to produce a new multimedia paradigm. However, their increased processing load on transmission servers handling large amounts of video and high-definition video must be considered. Our previously presented multi-vision system enables many users to view multiple videos by virtue of fast stream joining algorithms. The system joins (combines) much video stream on a transmission server into a single video stream. Then it transmits the joined single video stream to various client devices including mobile terminals. Because of its single-stream transmission, the client device enables decoding of the received video stream without special mechanisms. Especially, to join multiple video streams with high performance and without prediction inconsistency, the encoder fixes the number of a non-zero quantized coefficient (called TotalCoeff) to a certain in earlier work. Because this scheme of the fixing TotalCoeff in the encoder phase obviates the decoding process of each stream and modification of the prediction relation in the joining phase, it is possible to implement a stream joiner with a lightweight process. However, image degradation occurs because this scheme adds and drops the quantized coefficient. We emphasize that the amount of image degradation (called TotalCoeff fixing error or simply error) is dependent on the original TotalCoeff before fixing TotalCoeff and the Quantization Parameter (QP) because the quantization process has zero-rounding capability. Then we consider, using QP, the TotalCoeff estimation model and the error estimation model created by adding and deleting the coefficient. Under these models, we propose a QP control scheme to decrease this degradation without sacrificing high speed in the joiner(1).

Various environmental sounds exist around us in our daily life. Recently, environmental sound recognition has drawn great attention for understanding our environment. However, because environmental sounds derive from multiple sound sources, it is difficult to recognize them accurately. If we were able to separate sound sources before sound recognition as a pre-process, then recognition would be easier and more accurate. We assume that monaural microphones are widely installed in mobile devices used as recording devices. This paper therefore presents a proposal for monaural sound source separation of environmental sounds. Two-phase clustering using non-negative matrix factorization (NMF) is proposed to separate monaural sound sources. In this proposal, the time-variant gain feature is used as an attribute of an environmental sound for more efficient sound separation. (C) 2012 Elsevier Ltd. All rights reserved.

As described in this paper, an encoder that generates drift-error-resilient coded bitstreams is proposed. The proposed drift-error-resilient bitstreams are generated using the TotalCoeff or the total amount of non-zero DCT coefficients to estimate the drift-error distortion. This estimated drift error distortion is then examined in the motion estimation process. Results show that our proposed method can decrease drift error effectively compared to the original encoder and that it can achieve comparable PSNR to a conventional drift-error-resilient encoder. The main advantage of the proposed method compared to a conventional drift-error-resilient encoder is that, comparable PSNR performance, lower complexity and memory usage can be achieved without implementing an open-loop transcoder at the encoder feedback loop. (C) 2012 Elsevier Ltd. All rights reserved.

High functionality and interactivity of video services are strongly desired. This paper presents a proposal of an encoding scheme and a fast joining scheme for interactive multi-vision systems. The scheme enables numerous users to access any view area interactively with any desired resolution. The joining means the combination of multiple streams into a single video stream. The server joins pre-encoded multiple streams into one bitstream according to a user request, and then transmits it simultaneously to many clients. This structure obviates complexity at the client terminal because the client terminal simply receives and decodes the single stream. However, if joining processing loads are high, then it is difficult to provide services to many users. This paper presents a proposal of a new encoding method with prediction restrictions, MB-line size insertion, and Macro Block (MB) line byte alignment for a fast and lightweight stream joiner. Results of experiments described in this paper underscore the proposed method's effectiveness(1).

This paper presents a proposal for music mapping considering melody piece transition for a browsing system to visualize mood distances among music. Especially, we use the Melodies Markov model proposed in an earlier report to calculate some Conditional Divergences and to express a music relationship on a 2-D map applying two dimension reduction methods: Principal Component Analysis and Hermitian Form Model. We also compare these combinations of distance measure and mapping method and analyze which is useful for visualization of the relevance among genres or the relevance among artists(1).

This paper presents a proposal for an efficient software CAVLC decoder architecture in H.264/AVC based on level length extraction (LLE). Especially, level-decoding in a CAVLC decoder is addressed. Its features are summarized in two parts: higher efficient pipeline processing of level decoding, and simultaneous multiple calculations of multiple level codes separated from level decoding loop using Single Instruction Multiple data (SIMD) instruction. The former is achieved by separating Level calculation from Level parsing based on the LLE scheme, and removing branch operations in the level decoding loop. These improve the pipeline-processing efficiency. The latter removes Level calculation from the level decoding loop, and uses multiple Level calculations based on SIMD instruction. The proposed schemes emphasize the software architecture. They are therefore applicable to general computers. Consequently, they can also be integrated with other CAVLC opimization schemes for CoeffToken, TotalZeros, and RunBefore syntax elements. Based on results of evaluation experiments, we confirmed that the improved pipeline processing achieved 22% faster decoding speed compared with the conventional method, which used only the LLE scheme. The SIMD-based Level calculation also achieved a 38% faster decoder than before by integrating with the former part(1).

We have proposed so far a new multivision video provisioning scheme that enables numerous multiple users to access any view area with any desired resolution interactively and simultaneously. This paper proposes an adaptive view-area and quality scaling while preserving quick response for view-area movement. This can adaptively avoid redundant coding bits in conjunction with the movement of the viewing area. Simulation experiments and our implementation show the feasibility and effectiveness of our proposal.

As video technology advances rapidly, the need for simultaneous presentation of multiple videos is expected to increase in future. This paper utilized the orthogonal experiment to analyze visual effects of video physical features on multiple video sequences. The result shows that video position, human face, and motion attract attention effectively.

This paper proposes music mapping considering melody transition for browsing system to visualize mood distances among tunes. Especially, we use the Melodies Markov model proposed in [1] to calculate Conditional Divergence and express music relationship on 2-D map applying two Multi-dimensional Scaling (MDS) methods(1).

Recently, many researchers focus on high-definition and multivision video provisioning. [1] and [2] proposed a free-view access of the high-definition video and panoramic using multiple partitioned video streams. While they did not consider high complexity on the client to decode and play the video, [3] presented an innovative stream joining method for lower-complexity client as well as lower server load. It is, however, not practically feasible to serve videos simultaneously to a larger amount of users. This paper presents encoding techniques for a fast and simple stream joining that can achieve simultaneous multivision video provisioning.

This paper proposes a method to construct the scene composition of an event, a place and a situation by learning appearance pattern of objects in them. We recognize objects inside an image, and extract representative patterns based by analyzing the appearance combination of object in a large amount of images. Preliminary experiment shows that the proposal can create knowledge of scene composition with no contradiction and gives unfamiliar insights for the scene(1).

Visual information of music makes users to understand it rapidly, and is very useful for a fast and efficient music browsing. In this paper, we propose an automatic music image generation that extracts features from lyrics such as an event and a place using the probabilistic parametric mixture model and the scene composition knowledge(1).

This paper presents an implementation of a location-based sound-enhanced landscape album. Users can enjoy scenery images in any designated location enhanced by environmental sound and a Life-like agent's comments. This system bases our environmental sound provisioning system, and newly proposes a hierarchical-structured sound composition scheme to achieve natural and inartificial sound. This system has been released to the public for smart-phone users(1).

This paper discusses a new paradigm of network service and in-network resource control: relational metric based control. The relational metrics indicate the closeness relationship between objects in the real world, and these objects could be people, locations, things, and content. Closeness is measured by using a fusion of online and physical sensing. We will describe the system model and discuss possible service applications of this technology.

Fast scan and detection of the end of each CAVLC code is needed for bitstream operations like transcoding and manipulating. However, they need a larger computational load due to sequential and context-adaptive nature of CAVLC. That is why that it is dominant process in en entire decoding process. This paper proposes a fast skip scheme of CAVLC level code. Simulation results at the end of this paper show that the proposal about 70% reduction of CAVLC level code skip, and outperforms a conventional method.

In this paper the techniques for developing an efficient method of eye gesture tracking are explored. The goal is to verify the possibility of building an effective eye gesture input method relying on built-in hardware of the smart phone platform. The achieved results will be used as a basis for researching the technologies for gaze tracking on a mobile device. Effectiveness is measured according to detection accuracy, robustness, and processing performance. The results show limited success in achieving the stated goals, especially with respect to processing performance and provide a solid foundation for future work. © 2012 IEEE.

This paper presents a proposal of a super-realistic environmental sound synthesizer for a location-based sound search system. The location-based sound search system enables a user to experience any environmental sound in any desired remote place. This synthesizer is achieved by spatial audio mixing considering real-world conditions such as geographical features and townscapes, as well as dynamic situations such as those of town events and weather. Furthermore, a new velocity vector-based clustering method is proposed to reduce the cost of audio processing on the server side, which achieves a lower number of composing and decomposing clusters that have a direct and strong impact on high-cost audio processing such as spatial audio mixing(1).

With the promising development of user assist services on consumer electronics portable devices, service applications must continuously monitor dynamically changeable data and status, and judge whether their service actions should be activated. These operations, performed in real-time, result in heavy resource consumption on the portable devices. This paper proposes an innovative way to reduce the consumption by aggregating data access timings requested from multiple applications. Simulation shows that the proposed method can provide the highest precision data while reducing resource consumption. This mechanism can be installed into an operation system on a mobile device, or can operate as a middleware on the device. This must be one of key technologies to reduce battery power consumption of CE devices(1).

With the spread of high-definition video devices, and with the advance of digital broadcasting systems and video portal services in the Internet domains, numerous high-definition video presentations become available in everyday life. Under these circumstances, the desire of people to enjoy video contents in new ways had grown. This paper proposes a new multivision video provisioning system that enables numerous users to access any view area with any desired resolution interactively. The basic idea is that the partial area (tile)-based pre-coded video streams are stored in advance. These multiple streams are joined dynamically and delivered based on the user's view-area position. In this paper, the details of H.264/AVC tile-based video coding and a fast tile stream joiner scheme are presented. Simulation experiments and our practical implementation demonstrate the feasibility and effectiveness of our proposal.

New communication services among nearby passing people are growing rapidly thanks to the advance of short-range wireless communication technologies and the widespread use of mobile devices that use them. Time-offset communication capability will engender a new service paradigm and promote new service emergence. This paper presents a proposal of a new distributed mobile cache system to cache data temporarily in a designated local area. The system is realized through collaborative consumer electronics mobile devices which use no infrastructure network. The system overview and its middleware on the mobile devices are explained. Furthermore, new relay data algorithms are proposed to provide lower-overhead communication and a more robust caching capability. Finally, results of simulation experiments underscore the effectiveness of the proposed control algorithm in terms of the communication cost and the length of the cache time(1).

Video transcoding is one of the most essential technologies in consumer electronics server and gateway products for a variable bitrate video streaming system. Especially, for a real-time streaming system, drift-error resilient coded bitstreams that do not require high complexity transcoding processes are desired. For the generation of drift-error resilient coded bitstreams at the encoder side, an investigation on the drift-error propagation characteristic in H.264/AVC by applying different motion compensation (MC) precisions are carried out in this paper. The results show that the drift-error propagation in half-pixel MC precision increases due to the usage of six-tap filter. For quarter-pixel MC precision, the drift-error propagation decreases. Besides that, we found out that the requantization parameter and MC precision only affects the amount of error in the requantized frame but amount of error propagation depends on the MC precision and sequence of the pixel value. It is also found that drift-error propagation ratio is defined at some point regardless of MC precisions(1).

This paper presents a proposal of a new video provisioning scheme that enables numerous multiple users to access any view area with any desired resolution interactively. The basis is that the tile-based encoded multiple streams are joined dynamically and delivered based on the user's view-area position. This paper details a tile-based video coding scheme and a fast tile stream joiner scheme. Simulation experiments and our practical implementation illustrate the feasibility and effectiveness of our proposed scheme.

This paper proposes an innovative way to reduce the consumption by aggregating resource accesses timings requested from multiple applications. The simulation shows that the proposed method can provide the highest precision information while keeping the resource consumption lower.

This paper proposes a new distributed mobile storage architecture to store information in a local area by collaborating end-terminals without an infrastructure network. This paper shows the effectiveness of the proposed control algorithm by the simulation experiment.

In this paper, for the generation of drift-error-resilient coded bitstreams at the encoder side, an investigation on the drift-error propagation in H.264 by applying different motion estimation precisions is carried out. The results show that the drift-error propagation in half-pixel motion estimation precision increases due to the usage of six-tap filter. Mathematical model for estimating drift-error propagation is presented as well.

This paper proposes an on-demand soundscape generation and provisioning for a user to experience a real world in a requested remote place. This generation is achieved by spatial audio mixing considering a real world condition like geographical features or townscapes as well as dynamic situation such as town events or weather. The proposed velocity vector-based clustering can reduce the cost of composing/decomposing clusters of many sound sources.

The recent enhancement of mobile devices and wireless networks has enabled content services in mobile environments. Demand prediction is a traditional but powerful technique used for content services. However, it is hard to predict local demand in mobile environments because it depends not only on just user preference and the popularity of common content but also on other factors
users request content related to their locations
moreover, they are interested in the content uploaded by their friends who visited the location before them. Thus, we need to consider the context with multiple factors affecting the demand. In addition, we also need to consider the sequence of contexts.We call a sequence of contexts social context. This makes it difficult to predict local demand from users. In this paper, we propose a novel demand prediction engine that extracts local demand depending on social context and estimates what content will be requested there. To extract the demand, we use a log database and a pattern-matching technique in our prediction engine. To validate our prediction engine, we apply a prefetching service using the engine to a mobile content service. © 2011 IEEE.

It can be essential in the new-generation content services to predict the potential demands of people, which they themselves have not recognized or cannot express precisely. Social graphs representing the relationships between people are used for predicting demand in current Internet-based services. However, these graphs cannot represent the relationships of two users residing in common communities or common places. We propose representing not only a person but also things like communities and social events together as a single node in a social graph. This representation allows us to estimate who shares potential interests with a given person. We evaluated the estimation accuracy of our representation using an actual relational dataset from an academic database. Results show that our representation can estimate if two people share common interests that cannot be found with conventional methods that only use human nodes for estimation, and it can estimate the relations without using human nodes.

This paper proposes a new video delivery system with a lightweight, smooth, and quick-response accessibility to video scenes without interference by unstable mobile network conditions and also without consuming mobile device storage. This can avoid a video scene access delay brought on by mobile-network-specific delay and instability. We proposed an innovative streaming technology that consists of a view-context-aware asynchronous prefetching and streaming scheme and a seamless media assembly and media synchronization scheme. As an additional powerful extension for this system, we present a hierarchical prefetching asynchronous streaming technology that considers a video-scene structure. This can provide access to a general view through the whole content as well as to a detailed view around playback position by consideration of the video-scene structure and the user's viewing position. The implementation details on mobile phones and preliminary evaluation results are described, and its feasibility is also shown(1).

An efficient mobile video delivery with a light-weight, smooth and quick-responsible accessibility to video scenes has been proposed. This can avoid a content access delay brought by mobile-network-specific delay and instability. As an additional extension for our previous proposal, we present an innovative efficient access scheme with a hierarchical pre-fetching asynchronous streaming technology considering video scene structure. This can provide an access to a general view through a whole content as well as a detail view around a playback position by considering video scene structure and user's viewing position. The implementation details on mobile phones are also described at the end of this paper.

With the promising user assist services on portable devices, service controllers in multiple applications must continuously monitor dynamically changeable data and status, and judge whether its service action should be activated. These operations in a real-time manner bring heavy resource consumption on the portable devices. This paper proposes an innovative way to reduce the consumption by aggregating resource accesses timings requested from multiple applications. The simulation shows that the proposed method can provide the highest precision information while keeping the resource consumption lower. ©2009 IEEE.

With the promising user assist services on portable devices, service controllers in multiple applications must continuously monitor dynamically changeable data and status, and judge whether its service action should be activated. These operations in a real-time manner bring heavy resource consumption on the portable devices. This paper proposes an innovative way to reduce the consumption by aggregating resource accesses timings requested from multiple applications. The simulation shows that the proposed method can provide the highest precision information while keeping the resource consumption lower.

A light-weight, smooth, and quick-responsible accessibility to contents are truly needed for a practical usage of the mobile video services. We research on and develop a new innovative mobile video technology, which utilizes pre-download, pre-fetching and asynchronous network streaming schemes. This paper describes its innovative mechanisms, especially an asynchronous streaming technique and an asynchronous thumbnail data pre-fetching technique. Two prototype implementations of mobile video client are described in detail, and performance evaluations are given at the end of this paper.

Innovative Ajax-like functionality in a mobile video delivery system is introduced, where light-weight, smooth, and quick-responsible accessibility to contents can be provided. This technology consists of viewers' view context aware asynchronous pre-fetching and streaming scheme, and seamless media assembly and media synchronization scheme. This paper describes its innovative mechanisms and elemental technologies in detail. A prototype implementation of a mobile video client is described in detail, and performance evaluations are given at the end of this paper.

Personalized services can provide significant user benefits since they adapt their behavior to better support the user. Personalized services use a variety of data related to the user to decide their behavior. Thus personalized service needs a provisioning system that can collect the data that impacts service behavior and allows selection of the most appropriate service. However, in the coming ubiquitous environment, some data necessary for determining service behavior might be unavailable due to two possible reasons. One is that the data does not exit. The other is that the data exists but cannot be accessed. For example, users do not want to disclose their personal information, and service providers do not also want to expose data related to their knowhow in services. This paper describes a new service provisioning system for distributed personalization with private data protection. Specifically, the system selects applicable services by assessing how well each candidate service behaves when some data is missing. It then executes those selected services while hiding the users' and providers' private data in a distributed manner. We first summarize the requirements for a personalized service system, and introduce our fundamental policies for the system. The two main components of our system are then described in detail. One component is a service assessment mechanism that can judge if a service can work without data that can be used for adaptation. The second component is a service execution mechanism that can utilize private data while still ensuring privacy. This component divides service logic and executes divided logic where necessary data is available. The paper finally describes our prototype implementation and its performance evaluation results. (C) 2007 Elsevier Inc. All rights reserved.

We propose a scalable service execution control method for a service that is triggered when the context satisfies an execution condition. We call such services context-dependent. A context-dependent service needs to monitor the context that affects the execution condition all the time. The proposed method tremendously reduces the cost of context monitoring operations while it surely executes appropriate services when their execution conditions are satisfied.
Our method assumes that there are a lot of terminals monitoring contexts and a server that collects context information. from the terminals to manage service executions for a large number of services. Each context is modeled by a probabilistic time series, and thus, the probability that an execution condition is satisfied can be calculated at a given tine. The time at which the execution condition is likely to be satisfied is anticipated, and the server instructs relevant terminals to monitor contexts and to send their values to it before the anticipated execution. The time to collect context values is decided by the server; which considers the effect of the statistical variance of each context on the probability of satisfying the execution condition. Moreover, each terminal monitors its context and notifies the server when its value is out of the designated range, even if this occurs before the time set by the server. The range of values is set by the server: We conducted simulation experiments that showed the effectiveness of the proposed method.

Personalized services can provide significant user benefits since they adapt their behavior to better support the user. Personalized services use a variety of data related to the user to decide their behavior. Thus personalized service needs a provisioning system that can collect the data that impacts service behavior and allows selection of the most appropriate service. However, in the coming ubiquitous environment, some data necessary for determining service behavior might be unavailable due to two possible reasons. One is that the data does not exit. The other is that the data exists but can not be accessed. For example, users do not want to disclose their personal information, and service providers do not also want to expose data related to their know-how in services.
This paper describes a new service provisioning system for distributed personalization with private data protection. Specifically, the system selects applicable services by assessing how well each candidate service behaves when some data is missing. It then executes those selected services while hiding the users' and providers' private data in a distributed manner. We first summarize the requirements for a personalized service system, and introduce our fundamental policies for the system. The two main components of our system are then described in detail. One component is a service assessment mechanism that can judge if a service can work without data that can be used for adaptation. The second component is a service execution mechanism that can utilize private data while still ensuring privacy. This component divides service logic and executes divided logic where necessary data is available. The paper finally describes our prototype implementation and its performance evaluation results.

Personalized services can provide significant user benefits since they adapt their behavior to better support the user. Personalized services use a variety of data related to the user to decide their behavior. Thus personalized service needs a provisioning system that can collect the data that impacts service behavior and allows selection of the most appropriate service. However, in the coming ubiquitous environment, some data necessary for determining service behavior might be unavailable due to two possible reasons. One is that the data does not exit. The other is that the data exists but can not be accessed. For example, users do not want to disclose their personal information, and service providers do not also want to expose data related to their know-how in services.
This paper describes a new service provisioning system for distributed personalization with private data protection. Specifically, the system selects applicable services by assessing how well each candidate service behaves when some data is missing. It then executes those selected services while hiding the users' and providers' private data in a distributed manner. We first summarize the requirements for a personalized service system, and introduce our fundamental policies for the system. The two main components of our system are then described in detail. One component is a service assessment mechanism that can judge if a service can work without data that can be used for adaptation. The second component is a service execution mechanism that can utilize private data while still ensuring privacy. This component divides service logic and executes divided logic where necessary data is available. The paper finally describes our prototype implementation and its performance evaluation results.

This paper proposes a video transcoder that realizes efficiently the function of multirate conversion to multiple hitrates. The multirate conversion function is realized as follows. The decoder and the inverse quantizer are shared as a single processor. The requantizer and the reencoder with the same requantization step are shared among outputs. With these provisions, an architecture is proposed whereby the processing complexity in multirate conversion can be reduced. The number of requantization processing operations per macro-block is examined theoretically, and it is shown that the conversion processing complexity can be formulated as a function of the output bitrate. Lastly, the theoretical evaluation and the evaluation experiment based on these investigations show that the sharing of the requantizer/reencoder is the scheme that reduces the increase of the processing complexity as a function of the number of outputs, and can cope with any number of outputs with two to four times the processing complexity of the single-rate output scheme. © 2004 Wiley Periodicals, Inc.

A variety of information communication devices have been developed to help people maintain his/her awareness of interesting information. Notification systems play an important role in delivering up-to-date, important information via such devices. We have been developing a novel push-style notification system that delivers a succession of new messages onto the mobile terminal display in a well-timed manner. This system increases the probability of the user discovering interesting information from among a stream of items shown one after another, and accessing the network to get more details if desired. Needless to say, the appropriateness of the message to he notified is important to the user. To make a push-style notification system more useful and comfortable for users, however, it is also essential to deliver the information without interrupting the user's on-going task. Furthermore, some mechanism that can recover from the user's failure to notice a message is also needed. We fold all these abilities into the one term of "Message Reachability", and propose a comprehensive and flexible push-style notification system. One of its main features is a Bayesian "Notification User Model" which takes task state, user state and user characteristic into consideration. The other is an adaptive notification control mechanism based on the user model, and a compensation mechanism to handle messages that the user missed. This paper details the user model and message control mechanisms, and concludes by showing implementation details.

Protecting personal privacy is already seen as a crucial requirement in the implementation of service provisioning in the ubiquitous environment. From the view point of preserving personal privacy, the simplest approach would be for users not to reveal any kind of private information at any time while keeping the number of available services unrestricted. Meanwhile, from the service provider's point of view, though this has been not clearly stated so far, their service logics should also be hidden from others because those logics may leak their know-how. This paper presents a novel ubiquitous service provisioning mechanism that gives more opportunities for users to get available services while preserving the secrecy of users' and providers' sensitive information. The basic idea of this mechanism is to share service execution procedures between the service provider and the user by exchanging converted service logic described in the form of Bayesian decision networks. This paper describes the proposed mechanism and the conversion algorithm for the Bayesian networks, and details the system architecture and implementation.

This paper considers the hierarchical video coding system, and proposes a bit allocation rate control method by dynamic control of the hierarchical bit allocation ratio. As the first step, the MPEG-2 SNR scalability method is considered as the hierarchical video coding, and the relation between the hierarchical bit allocation and the basic/enhancement layer quality is analyzed. It is shown that the intermediate bit allocation ratio may degrade greatly the quality in the enhancement layer. The dependency of the image quality on the hierarchical bit allocation is noted, and video coding bit allocation control is proposed, where the hierarchical loss can be suppressed by dynamically controlling the bit allocation ratio for each picture. A control method is proposed particularly, where the optimal bit allocation ratio is calculated so that the characteristic function is maximized under the constraint that the bit allocation ratio for each GOP is kept constant. Lastly, a simulation experiment is presented. It is shown that the hierarchical loss, which is generated in the conventional bit allocation control, can be suppressed by the proposed control method, and the effectiveness of the proposed method is demonstrated.

In this paper, we propose an output packet remultiplexing scheme for an MPEG-2 multiprogram transport stream transcoder. The proposed scheme avoids STD buffer failure by dividing the packets into Program Specific Information (PSI), audio, and video and implementing remultiplexing. Specifically, PSI packets are multiplexed at a constant time interval, audio packets are multiplexed at the same times as the input times of the audio TS packets, and video packets are multiplexed so that the outputs of the video TS packets are at equal intervals. Finally, we clearly demonstrate through simulation experiments the ability to prevent STD buffer failure and the effectiveness of the proposed scheme. (C) 2002 Wiley Periodicals, Inc.

Today, many audiovisual delivery systems, including video streaming and video conferencing, are being developed for use over a range of networking technologies, the differing characteristics of which pose problems for service level interoperability. Multimedia transcoding is one means to provide interoperability between different types of audiovisual terminals and between terminals that connect to different networks. In this paper, we will present a multimedia transcoder system, which provides interoperability between video conferencing terminals on IP networks and mobile terminals on mobile networks.

As an appliance that achieves rate reduction of MPEG-2 multi-program transport streams, MPEG-2 multi-program transport stream transcoder has been proposed. This transcoder needs to multiplex packets without any STD buffer failure. However, no efficient packet-multiplexing scheme has been reported until now. In this paper, we propose packet-multiplexing scheme that no STD buffer failure occurs in MPEG-2 multi-program transport stream transcoder. This proposed scheme is implemented by the following ways: 1) PSI packets are output in constant interval, 2) audio packets are output in the same PCR as that of input transport stream, 3) video packets are output so that the intervals of each output packet for every video elementary stream should be constant. From simulation experiments, we show the status of STD buffer and show the effectiveness for our proposed scheme.

Today, many audiovisual delivery systems, including video streaming and videoconferencing, are being developed for use over a range of networking technologies, the differing characteristics of which pose problems for service level interoperability. Multimedia transcoding is one means to provide interoperability between different types of audiovisual terminals and between terminals that connect to different networks. This paper describes a flexible multimedia transcoder that enables service interoperability between different types of terminals across heterogeneous networks. In particular, the target of this development is to connect to H.323 client and OdBits servers from a client application in mobile terminals via the developed multimedia transcoder.

In this paper, we propose a video rate control scheme for MPEG-2 multi-program transport stream(TS)transcoder. At first, we explain its algorithm that considers the composition, rate distortion characteristics, and start-end points of pictures in current transport stream to be transcoded. Next, we show the transcoder with the proposed video control scheme, and explain the process flow. Finally, from simulation experiments, we compare the proposed rate control scheme to the conventional rate control scheme in terms of changes of a ratio between input bits and output bits, coding bits and picture quality.

In this paper, we propose a scalable video transmission system based on transcoding. First, we focus on the transcoder to reduce the bitrate of MPEG-2 bitstreams. We show that it is realized the layered coding by using the differential data between input and output of a transcoder. Next, we show the coding algorithm of this differential data. This algorithm can improve the coding efficiency by removing unnecessary code for differential data. By simulation experiments, we show that the proposed coding algorithm prevents the loss of coding efficiency in traditional hierachical coding, and show the effectiveness of proposed scheme.

Multimedia transcoding technology can provide interoperability between different types of audiovisual terminals and between terminals that connect to different networks. This paper describes a flexible multimedia transcoder that enables service interoperability between different types of terminals across heterogeneous networks.

In this paper, we focus on the MPEG-2 transcoder to provide the flexible video transmission by converting the compressed video bit streams, and study on the new coding scheme for realizing bit stream scalability by preserving and re-using the differenatial data generated by transcoding process. First, we show problems in the scalability by encoding and decoding process, and propose the scalable coding method by separating and composing of MPEG-2 bit streams. Especially, we aim at the improvement of coding efficiency, we consider the separating and composing algorithm. Next, we point that it is possible to estimate the changing of DCT coefficient block that generated by re-quantization process, and show the coding algorithm of this changing data by classificating coefficient value and re-constructing method of original DCT block. Last, by simulation experiments, we show that the proposed method prevents the increase of total bit-count by separating process, moreover it brings re-compressing effect. And by cascading the bit stream separating processor, it can easily realize the multi layer scalabirity, we show the effectivensss of proposed scalable coding method.

In this paper, we focus on the video transcoder as a bit rate reducer in a network node and propose a rate control scheme for a low-delay MPEG-2[1] video transcoder. First, we summarize the requirements of a rate control algorithm for low delay transcoding. Next, based on these requirements, we describe the proposed rate control scheme in detail. Then, we analyze the input and output buffer delay and calculate the total delay time of the proposed transcoder. Finally, we evaluate the proposed scheme from the simulation results of experiments on picture quality, transcoding delay time and GOP (Group Of Pictures) structure information (N/M value) of the input bit stream. Consequently, we showed that the proposed scheme can provide the same picture quality as a traditional scheme and is independent of the GOP structure.

In this paper, we focus on the video transcoder as a bit rate reducer in a network node and propose a rate control scheme for a low-delay MPEG-2(1) video transcoder. First, we summarize the requirements of a rate control algorithm for low-delay transcoding. Next, based on these requirements, we describe the proposed rate control scheme in detail. Then, we analyze the input and output buffer delay and calculate the total delay time of the proposed transcoder. Finally, we evaluate the proposed scheme from the simulation results of experiments on picture quality, transcoding delay time and GOP (Group Of Pictures) structure information (N/M value) of the input bit stream. Consequently, we showed that the proposed scheme can provide the same picture quality as a traditional scheme and is independent of the GOP structure.

As the transform method of video bit stream to the required bit stream format, the video transcoder has been adopted in various applications. In this paper, we focus on a MPEG video transcoder that can reduce the required bit rate by using re-quantization in the DCT domain. Furthermore, we propose a rate control method for this MPEG video transcoder that takes into consideration the estimation of drift-error propagation. First, we show the transcoder architecture with calculation of "Drift-Error Propagation Ratio". Next, we explain the proposed rate control method that uses minimization of the sum of re-quantization Error and drift-error. Finally, from simulation experiments, me compare the proposed rate control method to the traditional rate control method in terms of complexity, required buffer size and picture quality.

In this paper, we focus on the video transcoder as a bit rate reducer in a network node and propose a rate control scheme for a low-delay MPEG-2(1) video transcoder. First, we summarize the requirements of a rate control algorithm for low-delay transcoding. Next, based on these requirements, we describe the proposed rate control scheme in detail. Then, we analyze the input and output buffer delay and calculate the total delay time of the proposed transcoder. Finally, we evaluate the proposed scheme from the simulation results of experiments on picture quality, transcoding delay time and GOP (Group Of Pictures) structure information (N/M value) of the input bit stream. Consequently, we showed that the proposed scheme can provide the same picture quality as a traditional scheme and is independent of the GOP structure.

As the transform method of video bit stream to the required bit stream format, the video transcoder has been adopted in various applications. In this paper, we focus on a MPEG video transcoder that carl reduce the required lit rate by using re-quantization in the DCT domain. Furthermore, we propose a rate control method for this MPEG video transcoder that takes into consideration the estimation of drift-error propagation. Next, from simulation experiments, we compare the proposed rate control method to the traditional rate control method in terms of complexity, required buffer size and picture quality.

In this paper, we propose rate control algorithm based on dynamic quantizer matrix changing method to prevent VBV buffer over flow or under flow and rate control failure. We enconded and transcoded some MPE-2 test sequences with both MPEG-2 TM5 based rate control method and proposal method in 1.2 Mbps, and confirmed that proposal rate control algorithm is effective in VBV buffer over/under flow prevention and rate control failure prevention.

In this paper, we focus on the MPEG video transcoder for the bitrate reducer, and propose the rate control scheme considering re-quantization error. First, we state re-quantization process in the transcoder and mention that the transcoder has a different property from the encoder. Next, we investigate the re-quantization property in terms of step size, distortion and rate, and derive the rate-distortion property of re-quantization. Next, we propose the rate control scheme considring the rate-distortion of re-quantization with minimization of distortion generated by re-quantization in the transcoder. Last, we evaluate the proposed rate control scheme from the image quality, and show the effectiveness of proposed scheme.

In this paper, we propose the rate control scheme for low delay MPEG-2 video transcoder which reduce the bitrate of MPEG-2 bitstream in network node. At first, we state the requirement of the rate control for low delay transcoder. Next the rate control scheme for low delay transcoder is proposed. And, we analize the required delay time for the proposed scheme. At last, we evaluate the proposed scheme from the image quality, delay time and GOP-structure independency. As a result, we show the effectiveness of proposed scheme.

Multimedia communication services based on standard information, such as MPEG-1, MPEG-2, H.261 and H.263, had been indicated using multimedia scalability packages, which can provide multimedia information according to functionality and availability for users. In this paper, we propose the coding method of video information architecture in these packages, which are called "MSP". The MSP data architecture is defined using hierarchical architectures. We especially considered temporal feature information. The temporal scalability concept is described by both of MSP data and scalability content data. And then, as coding schemes, we attention to video hybrid scalability using spatial scalability and temporal one and scalability profiles are shown. Finally, we explained the new video temporal function by MSP data description.

We had indicated communication services using multimedia scalability pachage : MSP, which can provide multimedia information accoding to functionality and availability for users. This paper defines data structures of MSP using scalability architectures. Especially, when we made the hierarchical architecurue, we considered temporal feature information. And then, as coding schemes, we attention to video hybrid scalability using spatical scalability and temporal one and scalability profiles are shown.

We propose the Multimedia Scalability Packages(MSP) as the packages which are structured by contents with multimedia scalability structure. In this paper, we study the playing processing method and the cditting processing method based on the scalable structure which contents with scalability and the relations among contents with scalability. In particular, we explain the data structure for scalable playing and editting. At last, we state the application to the interacitve video communication service.