Time filter

Source Type

Tzelepis C.,Information Technologies Institute ITI | Tzelepis C.,Queen Mary, University of London | Ma Z.,Carnegie Mellon University | Mezaris V.,Information Technologies Institute ITI | And 5 more authors.
Image and Vision Computing

Research on event-based processing and analysis of media is receiving an increasing attention from the scientific community due to its relevance for an abundance of applications, from consumer video management and video surveillance to lifelogging and social media. Events have the ability to semantically encode relationships of different informational modalities, such as visual-audio-text, time, involved agents and objects, with the spatio-temporal component of events being a key feature for contextual analysis. This unveils an enormous potential for exploiting new information sources and opening new research directions. In this paper, we survey the existing literature in this field. We extensively review the employed conceptualization of the notion of event in multimedia, the techniques for event representation and modeling, the feature representation and event inference approaches for the problems of event detection in audio, visual, and textual content. Furthermore, we review some key event-based multimedia applications, and various benchmarking activities that provide solid frameworks for measuring the performance of different event processing and analysis systems. We provide an in-depth discussion of the insights obtained from reviewing the literature and identify future directions and challenges. © 2016 Elsevier B.V. Source

Tzelepis C.,Information Technologies Institute ITI | Tzelepis C.,Queen Mary, University of London | Mezaris V.,Information Technologies Institute ITI | Patras I.,Queen Mary, University of London
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

In this paper, we propose an algorithm that learns from uncertain data and exploits related videos for the problem of event detection; related videos are those that are closely associated, though not fully depicting the event of interest. In particular, two extensions of the linear SVM with Gaussian Sample Uncertainty are presented, which (a) lead to non-linear decision boundaries and (b) incorporate related class samples in the optimization problem. The resulting learning methods are especially useful in problems where only a limited number of positive and related training observations are provided, e.g., for the 10Ex subtask of TRECVID MED, where only ten positive and five related samples are provided for the training of a complex event detector. Experimental results on the TRECVID MED 2014 dataset verify the effectiveness of the proposed methods. © Springer International Publishing Switzerland 2016. Source

Markatopoulou F.,Information Technologies Institute ITI | Markatopoulou F.,Queen Mary, University of London | Mezaris V.,Information Technologies Institute ITI | Patras I.,Queen Mary, University of London
Proceedings - International Conference on Image Processing, ICIP

In this paper we propose a cascade architecture that can be used to train and combine different visual descriptors (local binary, local non-binary and Deep Convolutional Neural Network-based) for video concept detection. The proposed architecture is computationally more efficient than typical state-of-the-art video concept detection systems, without affecting the detection accuracy. In addition, this work presents a detailed study on combining descriptors based on Deep Convolutional Neural Networks with other popular local descriptors, both within a cascade and when using different late-fusion schemes. We evaluate our methods on the extensive video dataset of the 2013 TRECVID Semantic Indexing Task. © 2015 IEEE. Source

Michailidis I.T.,Information Technologies Institute ITI | Baldi S.,Information Technologies Institute ITI | Baldi S.,Technical University of Delft | Pichler M.F.,University of Graz | And 2 more authors.
Applied Energy

Energy efficient passive designs and constructions have been extensively studied in the last decades as a way to improve the ability of a building to store thermal energy, increase its thermal mass, increase passive insulation and reduce heat losses. However, many studies show that passive thermal designs alone are not enough to fully exploit the potential for energy efficiency in buildings: in fact, harmonizing the active elements for indoor thermal comfort with the passive design of the building can lead to further improvements in both energy efficiency and comfort. These improvements can be achieved via the design of appropriate Building Optimization and Control (BOC) systems, a task which is more complex in high-inertia buildings than in conventional ones. This is because high thermal mass implies a high memory, so that wrong control decisions will have negative repercussions over long time horizons. The design of proactive control strategies with the capability of acting in advance of a future situation, rather than just reacting to current conditions, is of crucial importance for a full exploitation of the capabilities of a high-inertia building. This paper applies a simulation-assisted control methodology to a high-inertia building in Kassel, Germany. A simulation model of the building is used to proactively optimize, using both current and future information about the external weather condition and the building state, a combined criterion composed of the energy consumption and the thermal comfort index. Both extensive simulation as well as real-life experiments performed during the unstable German wintertime, demonstrate that the proposed approach can effectively deal with the complex dynamics arising from the high-inertia structure, providing proactive and intelligent decisions that no currently employed rule-based strategy can replicate. © 2015 Elsevier Ltd. Source

Tzelepis C.,Information Technologies Institute ITI | Tzelepis C.,Queen Mary, University of London | Galanopoulos D.,Information Technologies Institute ITI | Mezaris V.,Information Technologies Institute ITI | Patras I.,Queen Mary, University of London
Image and Vision Computing

In this work we deal with the problem of high-level event detection in video. Specifically, we study the challenging problems of i) learning to detect video events from solely a textual description of the event, without using any positive video examples, and ii) additionally exploiting very few positive training samples together with a small number of "related" videos. For learning only from an event's textual description, we first identify a general learning framework and then study the impact of different design choices for various stages of this framework. For additionally learning from example videos, when true positive training samples are scarce, we employ an extension of the Support Vector Machine that allows us to exploit "related" event videos by automatically introducing different weights for subsets of the videos in the overall training set. Experimental evaluations performed on the large-scale TRECVID MED 2014 video dataset provide insight on the effectiveness of the proposed methods. © 2015 Elsevier B.V. Source

Discover hidden collaborations