Advances in Multimedia and Network Information System Technologies

Following the successful proven concept of small and specialized exhibitions (so called showrooms) run by research institutions as windows to scientific innovation, the network BONITA (a project financed by the INTERREG IV B – Baltic Sea Region) extends the physical showrooms to virtual ones. While the basic idea of the physical showroom is to have an attractive exhibition area for demonstrating cutting edge-technologies in a tangible and accessible fashion and transmitting technological knowledge between science and a region, the main idea of the virtual showroom is to have centralized access to several exhibits located in different places resulting in distributed knowledge and bridging the gap between the physical and virtual world of museums and showrooms and between the expert and the visitor. The presentation of what is now technically feasible should be just one aspect of the showroom. It should also create a connection to what is technically imaginable, whereby the visionary aspects of the technology are communicated. The combination of tangible benefits and interdisciplinary visions for the future is an exceeding interesting one. Firstly it allows specific innovations to find their way to market more quickly, since they gain a higher profile and are in the public eye. On the other hand, long-term trends can also be created interactively and discussed within different target groups.

As it is known, to transform an analogue image into digital form it is necessary to undergo the processes of sampling and quantification. The first of them consists of downloading at defined intervals data from analogue image the second one approximates analogue levels of brightness due to the closest digital levels. Both processes are the reason of an errors formation. Those errors have significant influence on the fields of the digital images transformation, in which it is necessary to synchronize images. This problem becomes particularly significant when we use the images gained from two different sources (scanner, digital camera). Anyone who uses the terms concern to images’ transformation, knows that bad synchronization can lead to wrong results. In his article authors present the algorithm which eliminates problem of bad images adjustment. The paper features the method of determination of the rotation angle and axis based on computation of the Normalized Mean Square Error (NMSE) coefficient.

The objective of the presented study is to show that it is possible to effectively separate harmonic sounds from musical sound mixtures for the purpose of automatic sounds recognition, without any prior knowledge of the mixed instruments. It has also been shown that a properly trained ANN enables to reliably validate separation results of mixed musical instrument sounds, and the validation corresponds with subjective perception of the separated sounds quality. A comparison between the results obtained with the use of the ANN-based recognition, subjective evaluation of the separation performance and the energy-based evaluation is provided.

The paper describes options to use a single point surround microphone to determine a direction of the sound source localization (SSL). A “soundfield” microphone with four transducers (capsules) characterized by cardioid responses. A unique mechanical design of the transducer results in its omnidirectional response. The microphone enables 3D sound acquisition in so called A-format. Upon further processing of signal it is possible to determine e.g. a direction of a sound source within space. The conducted experiments prove that a simple calculation algorithm is in particular feasible for a real time operation, and application of the soundfield microphone significantly simplifies mechanical design of the SSL system.

A method of recognizing events connected to danger based on their acoustic representation through Support Vector Machine classification is presented. The method proposed is particularly useful in an automatic surveillance system. The set of 28 parameters used in the classifier consists of dedicated parameters and MPEG-7 features. Methods for parameter calculation are presented, as well as a design of SVM model used for classification. The performance of the classifier was tested on a set of 372 example sounds, yielding high accuracy.

In this paper a new concept is proposed for finding communities in a social network based on a mixed graph theoretic model of a standard and a bipartite graph. Compared to previous methods the introduced algorithm has the advantage of noise-tolerance and is applicable independently of the size of the clusters in the graph. The cluster core-mining method is based on a modified MST algorithm. Clustering incomplete data is done by using bipartite graphs and fuzzy membership functions.

In this chapter the system based on camera and multimedia projector enabling a user to control computer applications by dynamic hand gestures is presented. The main objective is to present the gesture recognition methodology which bases on representing hand movement trajectory by motion vectors analyzed using fuzzy rule-based inference. The approach was engineered in the system developed with J2SE and C++ / OpenCV technology. OpenCV was used for image processing and J2SE with jFuzzyLogic package for gesture interpretation. Results of fuzzy rule-based and fixed threshold-based gesture recognition effectiveness are provided. As an example of system usage the so-called Interactive Whiteboard application is presented. Details on the application engineered are provided in the context of fuzzy inference processing.

Similarly to text, video is hierarchically structured. The analogies of text and video structures are discussed. Then the juxtaposition is presented of two indexing processes, i.e. of text and video indexing based on the content analysis of their structure units. Several frameworks of automatic detection and categorisation of video shots and scenes reporting the sport events in a given discipline in TV sports news have already been proposed. It has been observed that many sport videos such as archery, diving, soccer, and tennis have repetitive structure patterns. In the tests performed using the Automatic Video Indexer AVI shots and then scenes have been detected in tested TV news videos. Experimental results show good performance of the scheme of video scene detection of a given sport discipline in TV sports news. The Automatic Video Indexer is a research project investigating tools and techniques of automatic video indexing for retrieval systems.

A concept, practical realization and applications of a passive acoustic radar for automatic localization and tracking of sound sources were presented in the paper. The device consists of the new kind of multichannel miniature sound intensity sensors and a group of digital signal processing algorithms. Contrary to active radars, it does not emit a scanning beam but after receiving surroundings sounds it provides information about the directions of incoming acoustical signals. Practical examinations of the sensitivity and accuracy of the developed radar were also presented and discussed. The sensitivity of the realized acoustic radar was examined in a free sound field. Several kinds of sound signals were used, such as: pure tone from 125 to 16000 Hz, one third octave band noise in the same frequency range and impulsive sounds. The obtained results for every kind of signal groups were presented and discussed. Results from experiments show that in some cases even a small value of the signal to noise ratio was sufficient to localize the sound source correctly. A video camera can be pointed automatically to the place where the detected acoustical source is localized. Hence, information about the sound event direction can be used for the automatic and remote control of the PTZ (Pan Tilt Zoom) cameras. The automatic and continuous tracking in real time of the selected sound source movement is also possible. The proposed solution can significantly improve the functionality of traditional surveillance monitoring systems.

An application of a multiframe SR (superresolution) algorithm applied to video monitoring is described. The video signal generated by various types of video cameras with different parameters and signal distortions which may be very problematic for superresolution algorithms. The paper focuses on disadvantages in video signal which occur in video surveillance systems. Especially motion estimation and its influence on superresolution effectiveness is analyzed. In proposed initial solution a proper frame shift estimation is shown. Tests of the proposed algorithm performed video frames from real surveillance system in which many described difficulties were found. Result image examples show image resolution enhancement with plate numbers. The improvement of image quality is discussed in reference to further plate recognition.

The chapter provides an overview of essential analyses and comparisons helpful in corporate human resources management based on social network approach. Several ideas, measurements, interpretations and evaluation methods are presented and discussed, in particular group detection, centrality degree, dynamic analysis, social concept networks.

In the chapter an approach to solve the problem of service matchmaking and discovery is discussed. By a service we mean a independent component which has specified inputs and outputs and some functional and non-functional features. In our approach to describe services ontologies are used. It helps us to include some semantics to the services. Process of service discovery is started when user’s request is translated into SLA contract. The problem of matchmaking and discovery can be defined as a problem of finding service which fulfil user’s requirements as much as possible. In this work we present a system of service discovery and then our contribution to this domain – rough set-based approach to solve considered problem. Some simple example which illustrates proposed approach is shown.

In currently used networks there are no self-protection or autonomous defending mechanisms. This situation leads to the spread of self-propagating malware, which causes even more dangerous, and significant threats i.e. Botnets. In the EFIPSANS project a new architecture that includes self-* functionalities is introduced. Self-defending functionality, using data mining approach detects and reacts to some of network threats.

User Adaptive Systems (UAS) are growing in past time along with expansion of intelligent ubiquitous systems to every piece of embedded device around us. These systems provide a many of interesting and useful services to support our day life needs in all areas include biomedical world of our bodies. Every such service need however to provide their results by some kind of interface. Therefore the need of a simple, intuitive and graphically attractive interface is much appreciated. Our chapter is focused to several areas of user interface design, user interface adaptivity and visualization. In all areas of developed system we implement secured ways.

In the following paper we present the UCP-Tree and a new algorithm called CPGrowth for continuous pattern mining. The UCP-Tree is an aggregation tree that stores common subsequences of input sequences in the same nodes. The characteristic feature of the CPGrowth algorithm is that it does not require transitional trees at the next recursion levels. Moreover, new sequences can be inserted into the UPC-Tree without rebuilding, which is a considerable advantage considering that Trajectory Data Warehouses store massive amounts of data. In this paper we compared the efficiency of the proposed index with one of the fastest continuous pattern mining algorithms.

The importance of network security is rapidly increasing as more and more business is conducted via these systems. The proposed honeynet system can be used to detect bots or malware based on the evaluation of events occurring within a computer network. A honeypot is a trap set to detect, deflect or in some manner counteract attempts at unauthorized access to information systems. A honeynet (a network consisting of 2 or more honeypots) is used for surveillance of larger or more diverse networks for which one honeypot may not be sufficient. Honeynets are fast emerging as an indispensible forensic tool for the analysis of malicious network traffic. Honeypots can be considered to be traps for hackers and intruders and are generally deployed complimentary to Intrusion Detection Systems (IDS) and Intrusion Prevention Systems (IPS) in a network. Proposed system would be capable of providing cures for new fatal viruses which have not yet been discovered by security firms.

Bayes’ rule is universally applied in artificial intelligence and especially in Bayes reasoning, Bayes networks, in decision–making, in generating rules for probabilistic knowledge bases. However, its application requires knowledge about a priori distribution of probability or probability density that frequently is not given. Then, to find at least an approximate solution to a problem, the uniform a priori distribution is used. Do we always have to use this distribution? The paper shows that it is not true. The uniform prior should only be used if there is no knowledge about the real distribution. If however, we possess certain qualitative knowledge, e.g. that the real distribution is the unimodal one, or that its expected value is less than 0.5, then we can use this knowledge and apply a priori distribution being the average distribution of all possible unimodal distributions, instead of the uniform distribution. As a result we will usually get better approximation of the problem solution and will avoid large approximation errors. The paper explains the concept of average distributions and shows how they can be determined with a special method of granulation diminution of elementary events and probability.

This chapter presents a system that uses visualization techniques of multidimensional vector spaces for product search, after processing a document corpus by means of LSI. Introduced algorithms are able to dynamically generate mappings of the most closely related items according to the ‘thematic context’ in the current session. A user navigation starts with required category of Internet shop product at the centre of the diagram, and then focuses on refining choices. Display capabilities include several novel adaptive features. Examples of the system applicability and future directions are also discussed.

This contribution presents a method of adaptive user profile creation, modification at the field of Web search systems by using query terms and weighted terms of retrieved documents. One of the essential parts of the method is heuristic-based significant terms selection from relevant documents, which relevance evaluation is assigned by the user. Created during retrieval cycles sub-profiles, represent user interests, and are used for user query modification. The experiments concerning adaptive user profile, as a personalization mechanism of Web search system, are presented and discussed.

Search engines are found the most powerful tools of information systems. However, in case of multilingual systems, they are rather oriented towards shallow techniques. This paper outlines our study into refining query language on the comparison basis of some search engines that utilize different translation models in order to propose a novel search strategy which significantly imposes the Web traffic optimization, in particular of the Deep, or Hidden Web. The framework proposed reveals inadequacies of the trans lingual information processing and provides the benefits for the user interacting with the information system. Our analysis of the context and the syntactic structure enables the user to retrieve sentences in its natural form in at least two languages.

The chapter discusses selected issues in the field of music information retrieval on the Internet. Differences between music information retrieval and text based information retrieval are indicated. Ways of extracting information from audio files are discussed. ISMIR (Internet Systems of Music Information Retrieval) based on MIDI files, algorithms for identification of audio files based on so called fingerprints and ways of updating audio file metadata are discussed. New algorithms for positioning results of music files searches are proposed and essential elements of contemporary Internet Music Information Retrieval Systems are indicated.

The goal of the paper is to assess the usefulness of various text similarity measures for the two layered Internet search. In that approach the first layer is a generic Internet search engine. The second layer enables the user to evaluate, reorganize, filter and personalize the results of first layer search. It is run on a local work station and can fully exploit the so called user dividend. Crucial for that stage is assessing text similarity between text segments. The papers discusses classical, statistic text similarity measures as well semantic, WordNet based semantic measures. The results of an experiment show, that without word disambiguation techniques the semantic approaches can not outperform statistic methods.

A comparative analysis of the two most popular open source web frameworks for Java platform is done. The aim of the paper is to present modern software environments designed for implementing web applications and also to make final recommendations about web framework, which should be used in developing web application. The subjects of the analysis are SpringMVC and JavaServer Faces. The solution of the problem relies upon theoretical analysis of available framework features and upon empirical studies on implemented application designed to support managing a pet clinic.

E-learning is a constantly developing manner of distant courses distribution. With a technological progress its effectiveness and usability controlling abilities has increased and should be measured in order to provide more efficient web solutions. In this paper, multi-modal, user tracking, web usability testing platform is presented. Presented platform provides Afterwards selected Moodle based e-learning implementation usability testing was performed revealing its drawbacks and misfunctionality.

Developed e-learning materials and their use in teaching object-oriented programming course are presented in this paper. Presentations, films, laboratory instruction, multimedia textbook and handbooks are available to students via e-learning system Moodle. The training quizzes, quizzes for laboratories and the examination are created by Moodle individually for each student. All quizzes use the questions of common databases. These quizzes are resolved remotely and they are automatically evaluated by the Moodle platform.

This chapter deals with evaluation of user activities and their participation in collaborative processes realized within the supporting virtual environment. The main goal behind proposed solution is to provide standalone package with all necessary functionalities to obtain data from examined virtual collaborative system in form of logs. For this purpose the following has been designed and implemented: a repository with predefined log format representing source historical data for analyzes; supporting middleware services for proposed analytical approaches; end-user tool for time-line based mirroring and analyses of user activities. These basic functionalities are extended with possibility to extract various summative statistics [10] about performed user activities and possibility to export data in predefined format (e.g. MS Excel) for analyses in third party tools as e.g. IBM SPSS Modeler. Described analytical framework has been designed, implemented and tested mainly within KP-Lab System that represents new interesting application in the domain of virtual environments or e-Learning systems. Proposed architecture was designed as generic platform with possibility to be integrated with other systems such as Moodle or Claroline in order to reflect different user practices.

This paper focuses on producing screencasts for e-learning purposes. Creating screencasts by students in order to document their work in completing a task may be a good alternative to traditional tests in asynchronous learning. Presented in this paper experiment concerns analyzing student screencasts by extracting the media metadata. Metadata such as video resolution, fps or audio sampling frequency are compared in order to find most popular parameters of screencast created by regular screencasters, in this case represented by students.

This paper introduces the specification and the modeling of an opened agent-oriented system. We are interested in the development of a learning platform that offers learners a support for their projects development and tools to support collaboration among distributed users and tomanage traces of all interactions related to learners belonging to a given group.We used the Aalaadin methodology in order to model our system.