Catálogo de publicaciones - libros

Ontologies provide an attractive basis for the representation of semantic information to be attached to multimedia information. However, the flexibility available to develop one’s own ontology or partially reuse an existing ontology means that human input is unavoidable in the process of creating the annotations. We address the issue of how to use ontology reasoning services and natural language generation to provide presentations of relevant ontology structures for human use.

The movement of large numbers of people is important in many situations, such as the evacuation of a building in an emergency. In this paper, pedestrian dynamics during the evacuation of large areas is simulated using a computational intelligent technique, based on Cellular Automata. The characteristic feature of the proposed model is that the crowd consists of independent parts rather than treated as homogeneous mass. The crowd behaviour is artificially formatted by the response of each of these parts to the rule according to which each pedestrian reaches one of the possible exits. Furthermore, an efficient graphical user interface has been developed, in order to study various hypotheses concerning the pedestrians’ activity features. Collisions among pedestrians have been encountered while collective effects prominent at crowd behaviour have been also realised during simulation. Finally, the presence of fixed as well as user-defined moveable obstacles has been taken into account.

In this paper, as an active sensor imaging technique, a new target density function in form of the range density function(RDF) is developed by Gabor transform which is called short time Fourier transform (SIFT). It is shown that Gabor theory, (STFT) can be used as approach to imaging by active sensors by transmitting a waveform which is a kernel for this transform. Then an alternative signal dimension reduction approach is proposed to the developed technique by taking advantage of Walsh functions.

The use of mobile agents for data fusion in wireless sensor networks has been recently proposed in the literature to answer the scalability problem of client/server model. In this article, we consider the problem of calculating a near-optimal route for a mobile agent that incrementally fuses the data as it vis-its the nodes in a distributed sensor network. The order of visited nodes affects not only the quality but also the overall cost of data fusion. Our proposed heu-ristic algorithm adapts methods usually applied in network design problems in the specific requirements of sensor networks. It suggests the optimal number of MAs that minimizes the overall data fusion cost and constructs near-optimal itineraries for each of them. The performance gain of our algorithm over alter-native approaches is demonstrated by a quantitative evaluation..

Knowledge could be considered as the most critical dual factor — input and output — in production and scientific processes. Strategic and influential use of Information and Communication Technologies (ICTs) - the so called Information Society (IS) — reveals vast potential in knowledge production and economic development. In our point of view, two are the key issues in effective ICTs exploitation: (a) social consensus about quantity, quality and processing mechanisms of public information and (b) transformation process path between the general e-work model and the g-work paradigm. In this context, the knowledge-based development codifies the positive face of IS, in contrast to personal data abuse. The concept of the “new democratic deficit” is introduced in order to move into the foreground the significant dialectic relationship between ICTs and society.

This paper proposes a web search algorithm, which aims to distinguish irrelevant information and to enhance the amount of the relevant information in respect to a user’s query. The proposed algorithm is based on the Ant Colony Optimization algorithm (ACO), employing in parallel document similarity issues from the field of information retrieval. Ant Colony Optimization algorithms were inspired through the observation of ant colonies. In our approach, ants are used as agents through Internet, which are capable of collecting information, calculating the content similarity in each visited node and generating routing paths through the web.

Intelligent Energy efficiency is an important research topic for ad-hoc Wireless Sensor Networks (WSN). Power saving makes it possible to guarantee basic levels of system performance, such as connectivity, throughput and delay, in the presence of both mobility-immobility and a large number of sensor nodes. A large variety of approaches for intelligent energy-efficient schemes have been proposed in the literature focusing on different performance metrics. This article presents a comprehensive survey of recent energy-efficient schemes in ad-hoc wireless sensor networks, the application of which increases nodes’ lifetime and thus, network connectivity and survivability.

The rapid use of Internet has led to the investigation of digital watermarking as a complementary technology to traditional protection mechanisms. Significant research efforts and review works presenting unifying characteristics of different methods have been reported for audio and image watermarking. In the context of video watermarking, though, there is a great deal of non-uniformity in presented approaches. The objective in this paper is to give an in-depth overview of different video watermarking techniques in order to single out the particularities of that field. Furthermore, the paper presents a benchmarking framework for objective video watermarking performance evaluation. We conclude that novel techniques need to be implemented and unexplored video-driven approaches have to be investigated.

A novel human video object watermarking scheme is proposed in this paper, providing copyright protection of semantic content. The proposed method focuses on the existence of face and body regions within an initial image, however other cases can also be addressed. Initially detection of human video objects is achieved using two adaptive 2-D Gaussian models, one for skin color distribution modelling and the other for body localization modelling. A watermark is then designed using invariant Hu moments of each human video object, and the watermark insertion procedure is driven by an iterative encryption module based on chaotic functions. Performance of the proposed object based secure watermarking system is tested under various signal distortions and known cryptanalytic attacks.

In this paper, a novel semi-fragile watermarking authentication scheme is proposed, which is based on two watermarks: the local watermark and the global watermark. The two watermarks complement each other. The local watermark is generated from the quantized ACs in 8×8 DCT block. The global watermark is produced by the lowest frequency subband in DWT-transformed image. In watermark embedding, the dither-modulation quantization rule is adopted to improve the scheme’s security. In watermark detection and authentication, a novel authentication principle is proposed, which obtains good authentication performances. Experimental results prove that this scheme can exactly detect and verify the tampered location against JPEG compression and some other attacks.