Abstract: Due to the increasing heterogeneity in the current multimedia landscape, the delivery of multimedia content has become an important issue today. This heterogeneity is not only reflected by a plethora of different usage environments, but also by the presence of multiple (scalable) coding formats. Therefore, format-independent adaptation engines have to be used within a multimedia delivery platform, which are able to adapt the multimedia content according to a certain usage environment, independent of the underlying coding format of the content. By relying on automatically created textual descriptions of the highlevel syntax of binary media resources, a format-independent adaptation engine can be build. MPEG-21 generic Bitstream Syntax Schema (gBS Schema) is a tool that is part of the MPEG-21 Multimedia Framework. It enables the use of generic Bitstream Syntax Descriptions (gBSDs), i.e., textual descriptions in XML, to steer the adaptation of a binary media resource, using format-independent adaptation logic. In this paper, we address the design and performance evaluation of a multimedia delivery platform that relies on gBS Schema-driven adaptation engines. This platform is called MuMiVA; it is a fully integrated, extensible platform for multimedia delivery in heterogeneous usage environments, using streaming technologies. To demonstrate the flexibility of our multimedia delivery platform, we discuss the functioning of two different applications (i.e., exploitation of temporal scalability and shot selection) applied to two different coding formats (i.e., MPEG-4 Visual and H.264/AVC). Keywords— Content adaptation, Content delivery, MPEG-21 gBS Schema, XML transformations.

Abstract: Many activities on the web are driven by high-level goals of users, such as “plan a trip” or “buy some product”. In this paper, we are interested in exploring the role and structure of users’ goals in web search. We want to gain insights into how users express goals, and how their goals can be represented in a semi-formal way. This paper presents results from an exploratory study that focused on analyzing selected search sessions from a search engine log. In a detailed example, we demonstrate how goal-oriented search can be represented and understood as a traversal of goal graphs. Finally, we provide some ideas on how to construct large-scale goal graphs in a semi-algorithmic, collaborative way. We conclude with a description of a series of challenges that we consider to be important for future research.

Abstract: An approach for quality-of-service based replica management in a proxy-to-proxy network is presented. Management decisions are based on a metrics called replication affinity. Replication affinity combines two other affinity metrics we call (1) placement affinity and (2) reallocation affinity. Placement affinity is used to find the best location to enable QoS based media delivery to future clients. Reallocation affinity is used to make media stream replacement decisions in case of insufficient storage space. Using our approach content replacements can be varied in the granularity. Granularity variations are possible for compressed videos by using multiple description coding (MDC). The effect of using replication affinity is evaluated by emulation experiments. For the experiments the network simulator NS-2 has been used.

Abstract: Active and passive replication are powerful techniques to improve the quality of multimedia streaming. Most systems follow either the active or the passive approach. A well known example for active replication are Content Distribution Networks [8] that replicate data to predefined static locations. In contrast to that, P2P file sharing networks [2, 1] use passive replication where identical content is usually provided by different peers. We suggest a system that combines both techniques using Proxy Affinity, Request Affinity and Replication Affinity considering user preferences, user behaviour, hardware resources and networks capabilities.

Abstract: With the advent of Web 2.0 technologies a new attitude towards processing contents in the Internet has emerged. Nowadays it is a lot easier to create, share and retrieve multimedia contents on the Web. However, with the increasing amount in contents retrieval becomes more challenging and often leads to inadequate search results. One main reason is that image clustering and retrieval approaches usually stick either solely to the images' low-level features or their user-generated tags (high-level features). However, this is frequently inappropriate since the "real" semantics of an image can only be derived from the combination of low-level and high-level features. Consequently, we investigated a more holistic view on image semantics based on a system called Imagesemantics. This system combines MPEG-7 descriptions for low-level content-based retrieval features and MPEG-7 keywords by a machine learning approach producing joined OWL rules. The rule base is used in Imagesemantics to improve retrieval results.

Abstract: Due to the increasing performance requirements of decoding H.264/AVC in HDTV or larger resolutions, new approaches are necessary to enable real-time processing. According to the current trend to parallel computation in all performance classes, decoding of AVC must be mapped to these architectures even though this is complicated by the increased complexity and many data dependencies in the codec. We propose and evaluate different ways of using multithreading to speed-up our .NET implemented decoder. While slice based approaches scale best, this is not a flexible approach because of the reliance on specially encoded streams. Functional partitioning and macroblock pipelining prove to be a good alternative for almost all evaluated videos.

Abstract: While there is a lot of hype around various concepts associated with the term Web 2.0 in industry, little academic research has so far been conducted on the implications of this new approach for the domain of education. Much of what goes by the name of Web 2.0 can, in fact, be regarded as new kinds of learning technologies, and can be utilised as such. This paper explains the background of Web 2.0, investigates the implications for knowledge transfer in general, and then discusses its particular use in eLearning contexts with the help of short scenarios. The main challenge in the future will be to maintain essential Web 2.0 attributes, such as trust, openness, voluntariness and self-organisation, when applying Web 2.0 tools in institutional contexts.

Abstract: MPEG-21 Digital Item Adaptation (DIA) allows for a media codec agnostic multimedia adaptation approach which enables the implementation of generic adaptation engines. However, DIA is optimized for static, server-based adaptation. In this paper we introduce novel mechanisms to extend the DIA approach towards dynamic and distributed scenarios. This facilitates the placement of generic adaptation nodes which perform media codec agnostic and dynamic adaptation anywhere along the content delivery path. To validate our work we implemented such an adaptation node and evaluate its performance.

Abstract: Today’s increasing variety of media data results in a great diversity of XMLbased metadata, which describes the media data on semantic or syntactic levels, in order to make it more accessible to the user. This metadata can be of considerable size, which leads to problems in streaming scenarios. Other than media data, XML metadata has no concept of “samples”, thus inhibiting streamed (and timed) processing, which is natural for media data. In order to address the challenges and requirements resulting from this situation, the concept of streaming instructions is introduced. In particular, streaming instructions address the problem of fragmenting metadata, associating media segments and metadata fragments, and streaming and processing them in a synchronized manner. This is achieved by enriching the metadata with additional attributes to describe media and XML properties. Alternatively, a style sheet approach provides the opportunity to dynamically set such streaming properties without actually modifying the XML description.

Abstract: Enabling transparent and augmented use of multimedia content across a wide range of networks and devices is still a challenging task within the multimedia research community. Within multimedia frameworks, content adaptation is the core concept to overcome this issue. Most media adaptation engines targeting Universal Multimedia Access (UMA) scale the content w.r.t. terminal capabilities and network resource constraints and do not sufficiently consider user preferences. This paper focuses on a hybrid recommender technique for configuring a cross-modal utility model that guides adaptation of multimedia content. This approach additionally considers the user environment as well as demographic user data which leads to a personalized and increased multimedia experience. Based on a related adaptation decision technique we show how it is possible to offer a personalized adaptation for the individual user. We present a detailed evaluation of the approach based on results earned by subjective tests.

Abstract: Content adaptation is an important issue of multimedia frameworks in order to achieve universal multimedia access (UMA), i.e., to enable consumption of multimedia content independently of the given resource limitations, terminal capabilities, and user preferences. The digital item adaptation (DIA) standard, one of the core specifications of the MPEG-21 framework, supports content adaptation considering a wide range of networks, devices, and user preferences. Most adaptive multimedia frameworks targeting the UMA vision do not consider utility aspects in their adaptation decisions. This paper focuses on a generic semantic-based audio–visual utility model for DIA that aims to enhance the multimedia experience for the user. Our proposed model is able to take the semantics and the perceptual features of the content as well as the users' specific utility aspects into account. Based on a detailed analysis of these constraints, we will show how the model reacts on individual input data. For choosing the best adaptation decision considering resource limitations on client and server sides as well as network characteristics, we evaluate four algorithms for performing this adaptation decision taking task. We will discuss results according to some use case scenarios.

Abstract: Providing transparent and augmented use of multimedia content across a wide range of networks and devices is still a challenging task within the multimedia research community. Multimedia adaptation was figured out as a core concept to overcome this issue. Most multimedia adaptation engines for providing Universal Multimedia Access (UMA) scale the content under consideration of terminal capabilities and resource constraints but do not really consider individual user preferences. This paper introduces an adaptive multimedia framework which offers the user a personalized content variation for satisfying his/her individual utility preferences.

Abstract: Time- or history-management systems are implemented in a lot of existing applications. The visited web pages in browsers, the undoing list in nearly every editor or the recorded differences in version control systems are only some examples where time-based information plays an important role. Every kind of information (no matter if we talk about the last headlines in the newspaper or the last user inputs on the workstation) without an explicit date description is nearly useless. Due this fact we are wondering why no global time axis for recording general time-based information is integrated in today’s operating systems. We will introduce a time model to record all kinds of user actions and general time-based events as well. As a second part we will analyze how the visual output system can profit from such a global time axis as well. Up to now it takes a great effort to implement animated user interfaces and so they are rarely found on today’s software market. With the global conjunction of time it is possible to generate animations automatically for every kind of information. We explore the ways of using the time dimension for information presentation (for timely and not timely information as well). We will evaluate these so-called 4D user interfaces1 and introduce a programming model to take advantage of the time-dimension in multiple ways.

Abstract: The visualization of temporal information should not be seen as a special case. A lot of applications take already advantage of the time factor (e.g. capturing user events) but nearly all of them implement this feature completely proprietary. So why do we not embed the time as a ”first class citizen” in today’s operating systems that every application can use time based operations in an unified way. Such an invention would not only improve and standardize the capturing of temporal events but it would be of benefit for a temporal visualization system too. Within a 3D visualization space and a global time axis we introduce a printf4D() 1 method. With this method it is possible to display images, videos, text strings or any other kinds of information in an automatically animated way. As a first proposal we demonstrate this function in a ”flow of information” metaphor where information-objects are not displayed all at once but in a flowing manner over a certain period of time. Additionally we will show that printf4D() is not limited to temporal data. It can be extended automatically to any kind of static information.

Abstract: Is Web 2.0 just hype or just a buzzword, which might disappear in the near future One way to find answers to these questions is to investigate the actual benefit of the Web 2.0 for real use cases. Within this contribution we study a very special aspect of the Web 2.0 the folksonomy and its use within self-directed learning. Guided by conceptual principles of emergent computing we point out methods, which might be able to let semantics emerge from folksonomies and discuss the effect of the results in self-directed learning.

Abstract: Folksonomies, collaboratively created sets of metadata, are becoming more and more important for organising information and knowledge of communites in the Web. While for a single user the difference to keyword assignment is marginal, the power of folksonomies emerges from the collaborative aspects. Folksonomies are already issue of research.Within this publication we analyse underlying statistical properties of broad folksonomies aiming to identify laws and characteristics, which allow inferring properties for folksonomy based retrieval. The actual benefit of folksonomies for retrieval and the derived methods are concluded from experiments with aggregated data from del.icio.us1.

Abstract: The evaluation of retrieval mechanisms for inter-method comparison is necessary in academic as well as in applied research. A major issue in every evaluation is in which way and to what extent the actual perception of the user from the target user group is integrated. Within multimedia retrieval systems the impressions and perceptions of users vary much more than in text retrieval. Empirical studies are a common tool in social science and offer a way to research the correlation between the user perception and the computed similarity between pairs of multimedia documents or a query and the set of results. This approach can be used to complement and extend current evaluation approaches. Within this contribution we summarize general methods from social science and psychology for the interested reader in the area of computer science with some knowledge about statistics. Furthermore we give two examples of undertaken empirical experiments and their outcomes. Within the first one the perception of users is investigated and compared to factors like background and gender, while in the second study metrics are tested upon their ability to reflect the notion of similarity of users. Both experiments aim to give examples and insight on how empirical studies can be used in multimedia research in general and multimedia retrieval evaluation in special.

Abstract: There is a number of image retrieval systems which allow a fast similarity search in large image databases. But to our knowledge there are no image retrieval systems which bring together information extraction from the image as well as object recognition and classification of the object analyzed. This paper introduces techniques to enable the extraction of botanical characteristics from visual features to support semi-automatic plant recognition. The identification of plants requires recognition and determination of plant species specific features such as bloom colour, inflorescences, shape of blooms, number of petals and shape of leaves. Our approach is to improve an existing medicinal plant database, called MedPhyt, by this basic requirement. We demonstrate the first steps towards a semiautomatic system for the identification of a plant species or at least plant families by the combination of both bloom colour and contour information under consideration of a specific content and knowledge domain using the features of MPEG-7.