Advanced Internet Based Systems and Applications

Text in video is a very compact and accurate clue for video indexing and summarization. This paper presents an algorithm regarding word group as a special symbol to detect, localize and extract video text using support vector machine (SVM) automatically. First, four sobel operators are applied to get the EM(edge map) of the video frame and the EM is segmented into N×2N size blocks. Then character features and characters group structure features are extracted to construct a 19-dimension feature vector. We use a pre-trained SVM to partition each block into two classes: text and non-text blocks. Secondly a dilatation-shrink process is employed to adjust the text position. Finally text regions are enhanced by multiple frame information. After binarization of enhanced text region, the text region with clean background is recognized by OCR software. Experimental results show that the proposed method can detect, localize, and extract video texts with high accuracy.

Due to the growing amount of digital media an increasing need to automatically categorize media such as music or pictures has been emerged. One of the metadata standards that has been established to search and retrieve media is MPEG-7. But it does not yet exist a query format that enables the user to query multimedia metadata databases. Therefore the MPEG committee decided to instantiate a call for proposal (N8220) for an MPEG-7 query format (MP7QF) framework and specified a set of requirements (N8219). This paper introduces a MP7QF framework and describes its main components and associated MP7QF XML schema types. The framework makes use of the MPEG-21 digital item declaration language (DIDL) for exchanging MP7QF Items along various MP7QF frameworks and client applications. An MP7QF Item acts as container for the input query format and output query format of a user query request.

This paper concentrates on components of the framework such as session management, service retrieval and its usability and excludes consciously definitions and explanations of the input and output query format.

In this paper we introduce a new approach to multimedia data semantic characterisation and in particular television programmes fingerprinting, based on multimodal content analysis and fuzzy clustering. The definition of the fingerprints can be seen as a

space transformation process

, which maps each programme description from the surrogate vector space to a new vector space, defined through a fuzzy clustering method. The fuzzy fingerprint model is well suited for similarity based information retrieval, and it captures ”semantic” similarities coming from common pattern in the programme data, at different semantic levels.

MPEG-7 (Multimedia Content Description Interface) is a standard for describing the content of multimedia data and adopts XML (eXtensible Markup Language) Schema as its Description Definition Language. However, as a forthcoming standard for querying XML data/documents, XQuery (XML Query Language) has weakness in the query of vector-based feature described in MPEG-7. Thus, we propose an extension to XQuery, that is, VeXQuery (Vector-based XQuery) language, to resolve the problem of vector-based feature query in MPEG-7 retrieval. To fulfill the vector-based feature query, VeXQuery gives a set of vector similarity measurement expressions and defines the formal semantics of VeXQuery in terms of XQuery itself. Through this approach VeXQuery can be integrated seamlessly into XQuery.

Defining and processing flexible queries is an important research topic in database area. We may distinguish two ways to support flexible querying in databases: (i) developing interface systems that allow queries in pseudo-natural languages, or (ii) developing an extended SQL-like languages. In this paper, we have adopted this second approach. The objective of this paper is to introduce a conceptual query language for accessing FSM-based databases. FSM is a data model that has been recently proposed.

In this paper, we focus on the managing of multimedia document and more precisely on the annotation and the generation of adaptable multimedia documents. Our solution is directed towards analysing the ways to “bridge the gap” between physical and semantic levels, for multimedia document modelling and querying. Our goal is to describe how to model and unify features elicited from content and structure mining. These descriptors are built from the various features elicited from the multimedia documents using available processing techniques. The personalization enables dynamic re-structuring and re-construction of hypermedia documents answering to the user queries. However, more factors should be considered in handling hypermedia documents. Once queried, documents can be adapted by using an indexing scheme, which exploits multiple structures. We can process queries efficiently with minimal storage overhead. We suggest for that, the adaptation of multimedia document content with user needs and preferences. This approach is based on the OOHDM methodology extension with the use of the metadata.

Besides XPath, XQuery is an XML query language developed by W3C. Within this paper, which is an extended version of [6], we present an approach for the optimization of the execution time of nested XQuery queries based on type information of the input XML document given in XML Schema. The optimization approach deletes unnecessary sub-expressions and shifts predicates to inner sub-expressions wherever possible in order to avoid the generation of unnecessary intermediate data. A performance analysis of our prototype demonstrates the optimization of the execution time of various current XQuery evaluators when using the proposed optimization approach.

Geographical (spatial) information about the real world changes rapidly with time. We can simply see examples of these changes when we look at any area. New buildings, new roads and highways, and many other new constructions are added or updated. Spatial changes can be categorized in two categories: (1) Discrete: changes of the geometries of physical entities (i.e., buildings) and (2) abstract: moving objects like airplanes, cars or even moving people. Spatio-temporal databases need to store information about spatial information and record their changes over time. The main goal our study in this paper is to find an efficient way to deal with spatio-temporal data, including the ability to store, retrieve, update, and query. We offer an approach for indexing and retrieving spatio-temporal data (AIRSTD). We concentrate on two main objectives: (1) Provide indexing structures for spatio-temporal data and (2) provide efficient algorithms to deal with these structures.

The power of geographic information system is to help managers make critical decisions they face daily. The ability to make sound decisions relies upon the availability of relevant information. Typically, spatial databases do not contain much information that could support the decision making process in all situations. To extend the available dataset, we propose an approach for the enrichment of geographical databases (GDB) and especially their semantic component. This enrichment is performed by providing knowledge extracted from web documents to supplement the aspatial data of the GDB. The knowledge extraction process is reached through the generation of condensed representation of the relevant information derived from a web corpus. This process is carried out in a distributed fashion that complies with the multi-agents paradigm.

Timestamping is a cryptographic technique providing us with a proof of existence of a message/document at a given time. Several timestamping schemes have already been proposed. In this paper, we shortly review existing schemes and then fully define a new timestamping system based on skip lists. We show that our scheme offers good performances.

Most existing end-to-end congestion control protocols employ packet loss or round-trip delay to imply network congestion. However, this kind of implicit signal mechanism may not work well in heterogeneous networks. Recently some router-assisted congestion control protocols are proposed to address this challenge and Quick Flow Control Protocol (QFCP) is one of them. QFCP allows flows to start with high initial sending rates and to converge to the fair-share rate quickly based on feedback from routers. The rate allocation algorithm is quite simple and only needs to be run periodically by routers. We have implemented QFCP in ns-2. Simulations have been done to address the issues such as flow completion time of Poisson-arriving Pareto-distributed-size flows, adaptability to changing flow numbers, fairness on flows with different RTTs, robustness to non-congestion packet losses, and performance on multiple bottleneck links. The preliminary results are promising.

The pervasive computing paradigm provides the user with a uniform computing space available everywhere, any time and in the most appropriate form and modality. These aspects produce the need for user interfaces that are usable, multimodal and personalized for each user. In this paper multimodality is discussed. In particular, features and computational issues of the multimodal interaction are analyzed in order to examine methodological aspects for the definition of multimodal interaction languages for pervasive applications. The multimodality is faced at a grammar level, rather than at a dialogue management level. This means that different unimodal inputs are considered as a unique multimodal input that is sent to the dialogue parser that uses the grammar specification to interpret it, rather than to be distinctly interpreted and then combined. The main objective of the paper is thus to explore multimodal interaction for pervasive applications through the use of a multimodal language rather than through the integration of several unimodal languages.

Outranking methods, a family of multicriteria analysis tools, cope better with the ordinal aspects of spatial decision problems. However, it is recognized that these methods are subject to computational limitations with respect to the number of alternatives. This paper proposes an approach to generate these alternatives based on a planar subdivision of the study area. The planar subdivision is obtained by combining a set of criteria maps. The result is a set of non-overlapping polygons/spatial units. Punctual, linear and areal decision alternatives, conventionally used in spatial multicriteria analysis, are then constructed as an individual, a collection of linearly adjacent or a collection of contiguous spatial units. This permits to reduce substantially the number of alternatives enabling the use of outranking methods.

One of the most successful and simplest Internet-based systems to promote the spread of knowledge are discussion forums. Unfortunately, not much research has been done to improve them by introducing adaptive capabilities. In this paper we discuss a market-oriented mechanism to promote the knowledge exchange activity in discussion forums. Our system provides a strategy to dynamically adapt the structure of the forums by using statistics about users behavior. By doing so, we minimize the response time in which questions are satisfactorily answered. In addition to that, the effort needed to generate useful information is also reduced.

The study proposed the method of keyword enhanced Web structure mining which combines the ideas of Web content mining with Web structure mining. The method was used to mine data on business competition among a group of DSLAM companies. Specifically, the keyword DSLAM was incorporated into queries that searched for co-links between pairs of company Websites. The resulting co-link matrix was analyzed using multidimensional scaling (MDS) to map business competition positions. The study shows that the proposed method improves upon the previous method of Web structure mining alone by producing a more accurate map of business competition in the DSLAM sector.

The Management Systems of Road Transport (MSRT) must include responsibility for planning the routes and schedules of vehicles fleet involved in the road haulage, distribution and logistics. It must ensure that all operations are carried out in maximum safety, environmental controls and traffic congestion, driver hours, customs requirements, and minimum cost. The complexity of the real-time scheduling of transport orders which comes in an asynchronous and dynamic way makes the MSRT especially suitable for using techniques from Distributed AI. To manage this complex field addressed under a high degree of dynamism and uncertainty which is characterized by an inherent distribution of knowledge and control, we propose in this work, a modeling of an MSRT by a multi-agents system, the modeling of the agents and their interaction by AUML language and we deal with the cooperation during tasks planning in the MSRT.

Organization modeling is recognized as an essential mechanism for structuring the design of Multi-Agent Systems and coordinating their executions. In this paper, an organizational model for Cooperative Information Gathering Systems (CIGS) is proposed. This model has three levels of abstraction, starting with a general system description and progressively transformed onto a concrete organization. This model is organized around reusable and conceptual basic bricks corresponding to roles and protocols. The roles and their participations in the protocols are defined according to a deontic structure inspired from the GAIA methodology, but specified formally using Petri Nets with Objects. The originalities and the advantages of our proposal are i) the definition of an organizational model adapted to the CIG ii) a formalization of the roles, which makes it possible to analyze and simulate them before their deployment iii) the possibility to reuse the components of the model.

Many algorithms in the literature have been targeting the problem of scheduling divisible workloads (those loads that are amenable to partitioning in any number of chunks). Unfortunately, such algorithms have a number of shortcomings such as the sole reliance in their computations on CPU speed, and the assumption that a definite set of workers are available and must participate in processing the load. These constraints limit the utility of such algorithms and make them impractical for a computing platform such as the Grid. In this paper, we propose an algorithm, MRRS, that overcomes these limitations and adopts a worker selection policy that aims at minimizing the execution time. The MRRS has been evaluated against other scheduling algorithms such as UMR and LP and showed better results.

An essential issue in peer-to-peer data management is to keep data highly available all the time. The paper presents a replication protocol that adjusts autonomously the number of replicas to deliver a configured data availability guarantee. The protocol is based on a Distributed Hash Table (DHT), measurement of peer online probability in the system, and adjustment of the number of replicas accordingly. The evaluation shows that we are able to maintain the requested data availability achieving near to optimal storage costs, independent of the number of replicas used during the initialization of the system.

We are interested in protein classification based on their primary structures. The goal is to automatically classify proteins sequences according to their families. This task goes through the extraction of a set of descriptors that we present to the supervised learning algorithms. There are many types of descriptors used in the literature. The most popular one is the n-gram. It corresponds to a series of characters of n-length. The standard approach of the n-grams consists in setting first the parameter n, extracting the corresponding ngrams descriptors, and in working with this value during the whole data mining process. In this paper, we propose an hierarchical approach to the n-grams construction. The goal is to obtain descriptors of varying length for a better characterization of the protein families. This approach tries to answer to the domain knowledge of the biologists. The patterns, which characterize the proteins’ family, have most of the time a various length. Our idea is to transpose the frequent itemsets extraction principle, mainly used for the association rule mining, in the n-grams extraction for protein classification context. The experimentation shows that the new approach is consistent with the biological reality and has the same accuracy of the standard approach.

Information systems’ activities are increasingly becoming distributed. Many systems, therefore, need to exchange documents in order to correctly perform a critical activity. A growing number of document exchange behavior creates new requirements, such as automatically matching document structure, simplifying user’s tasks in creating mappings, and automatically generating data transformation. In our work, we are interested in all these processes: matching, mapping, and data transformation. We propose, in this paper, to extend the XML schema matching used for document exchange to cover constraints management. We show how the constraints improve the performance of schema matching. We also propose XME (XML Mapping Expression) which is an expression model of mappings along with a collection of transformation operators. These extensions take part in our platform for document exchange (ASMADE) that automates matching and generates data transformation instances.

The data warehouse design raises several problems such as the integration of heterogeneous data sources. In fact, the main difficulty is how interpret automatically the semantic of the heterogeneous and autonomous data. In addition, in information system design, the use of ontology becomes more and more promising; it contributes to avoid semantic and structural ambiguities. This research introduces the concept of decisional ontology as an assistance tool for the specification of analytical requirements. For this, we propose an ontology based method to standardize the multidimensional terminology extracted from several heterogeneous data sources.

This paper presents an information retrieval methodology which uses Formal Concept Analysis in conjunction with semantics to provide contextual answers to Web queries The

conceptual

context defined in this article can be

global

- i.e. stable- or

instantaneous

- i.e. bounded by the global context. Our methodology consists first in a pre-treatment providing the global conceptual context and then in an online contextual processing of users’ requests, associated to an instantaneous context. Our information retrieval process is illustrated through experimentation results in the tourism domain. One interest of our approach is to perform a more relevant and refined information retrieval, closer to the users’ expectation.

There are many problems associated with requirements engineering. In fact, some concepts can be seen differently according to the user’s context. The requirement specification generates several problems of incoherence, ambiguity and difficulty when users admit different contexts. These problems may lead to poor requirements and unsatisfactory or unacceptable future system. For this reason, we propose a multi-representation ontology to solve these conflicts and problems. In this paper, we propose our motivation and we expose our top-down approach for building a multi-representation ontology dedicated to multi-context requirements specification in the e-learning field.

Given a large number of data sources, each of them being indexed by attributes from a predefined set

$\cal{A}$

and given a query

q

over a subset

Q

of

$\cal{A}$

with size

k

attributes, we are interested in identifying the set of all possible combinations of sources such that the union of their attributes covers

Q

. Each combination

c

may lead to a rewriting of

q

as a join over the sources in

c

. Furthermore, to limit redundancy and combinatorial explosion, we want the combination of sources to produce a

minimal

cover of

Q

. Although motivated by query rewriting in OpenXView [3], an XML data integration system with a large number of XML sources, we believe that the solutions provided in this paper apply to other scalable data integration schemes. In this paper we focus on the cases where the number of sources is very large, while the size of queries is small. We propose a novel algorithm for the computation of the set of minimal covers of a query and experimentally evaluate its performance.

This paper introduces a novel method for semantic annotation of web pages. We perform semantic annotation with regard to unwritten and empirically proven agreement between users and web designers using web patterns. This method is based on extraction of patterns, which are characteristic for a particular domain. A pattern provides formalization of the agreement and allows assigning semantics to parts of web pages. We will introduce experiments with this method and show its benefits for querying the web.

The accuracy of interface description is very important to service composition and dynamic selection of service-oriented systems. It is desirable to describe Web service formally so as to improve the ability of service orchestration. This paper presents a formal interface theory for specifying Web service by extending the existing with the ability to model interface behaviour with transactions at the levels of signature, conversation and protocol. Signature interface and conversation interface model the static invocation relations in Web service interfaces, and protocol interface describes the temporal invocation information. A formal semantics of protocol interface is presented. Based on the semantics, the protocol interface can be transformed into a Labeled Transition System (LTS). Additionally, the compatibility and substitutivity relation conditions between Web services are also proposed.

The hierarchical structure of a document plays an important role in understanding the relationships between its contents. However, such a structure is not always explicitly represented in web documents through available html hierarchical tags. Headings however, are usually differentiated from ‘normal’ text in a document in terms of presentation thus providing an implicit structure discernable by a human reader. As such, an important pre-processing step for applications that need to operate on the hierarchical level is to extract the implicitly represented hierarchal structure. In this paper, an algorithm for heading detection and heading level detection which makes use of various visual presentations is presented. Results of evaluating this algorithm are also reported.

The emergence of web services as a new technology supporting the future Web has motivated several researchers to investigate in this field. In addition, the possibility of selecting and integrating Web services, in or between organizations, is very useful for the improvement of their profitabilities. In such a case, we introduce the term of

web service composition

. In this paper, we propose an approach called WeSCo_CBR (Web Service Composition founded on Case Based Reasoning) that aims at enhancing the process of web service composition by using a case-based reasoning technique which originates from artificial intelligence. Furthermore, we regard efficient description and management of information semantics as a major requirement to enable semantic interoperability. We opt to integrate ontology so that we can apply a reasoning to help perform meaningful web service composition.

We describe a path approximate search process based on an extended editing distance designed to manage ‘don’t care characters’ (*) with variable length (*i~j) in a path matching scheme extending XPath. The structural path is bounded to conditional properties using variables whose values are retrieved thanks to a backtracking processed on the editing distance matrix. This system provides a dedicated iterator for a XML query and processing scripting language that features large XML document collection management, joint operations and extraction features.

Much research work has recently focused on the problem of representing historical information in XML. This paper describes a number of temporal XML data models and provides their comparison according to the following properties: time dimension (valid time, transaction time), support of temporal elements and attributes, querying possibilities, association to XML Schema/DTD, and influence on XML syntax. We conclude that the approaches of managing time information using XML mostly do not require changes of current standards.

In semi-structured information systems, generally, the adaptation of documents is essential to give the user the feeling that the query result is adapted to his preferences. The user’s needs can be defined in a user profile. But, in the literature, adaptation systems are designed for a particular domain and are oriented towards either navigation adaptation or content adaptation. Adaptation takes place after the user’s query has been evaluated. So, in this paper, we contribute to propose an adaptation algorithm which is domain independent and whose adaptation takes place before user’s query evaluation. This algorithm consists in enriching the user query on the basis of user profile in order to adapt the results to the user.

Mediators are critical for service scenarios, since providers and consumers frequently have opposite goals, e.g. to maximize profit with minimal resource (provider perspective) against to maximize satisfaction with minimal cost (consumer perspective). Nonetheless, semantic mediation is not explicitly taken into account in the Service Oriented Architecture (SOA). In this paper, it is presented an extension of SOA which includes the mediator as first-level entity, besides the client, the provider, and the registry of the services. The assignment of semantics to information that is processed by the mediator through ontologies, and the formalization of mediator’s behavior are included among the advantages of the proposal.