Grabner, HVan Gool, LVarvarigou, TKosmopoulos, DVeres, GVoulodimos, AModelling and classification of time series stemming from visual workflows is a very challenging problem due to the inherent complexity of the activity patterns involved and the difficulty in tracking moving targets. In this paper, we propose a framework for classification of visual tasks in industrial environments. We propose a novel method to automatically segment the input stream and to classify the resulting segments using prior knowledge and hidden Markov models (HMMs), combined through a genetic algorithm. We compare this method to an echo state network (ESN) approach, which is appropriate for general-purpose time-series classification. In addition, we explore the applicability of several fusion schemes for multicamera configuration in order to mitigate the problem of limited visibility and occlusions. The performance of the suggested approaches is evaluated on real-world visual behaviour scenarios. (C) 2011 Elsevier Ltd. All rights reserved.http://hdl.handle.net/123456789/21343ISI:0002951057000092486010.1016/j.neunet.2011.06.00185280893-6080engPERGAMON-ELSEVIER SCIENCE LTDNeural Networkstime series analysisartificial neural networkPrior knowledgeWorkflow monitoringonline monitoringLimited visibilitygenetic algorithmIndustryVisual tasksautomationActivity recognitionTracking moving targetsvisibilityworkflowTime series classificationsEcho state networksInherent complexityWorkflowclassificationOn-line classificationIndustrial environmentsarticleindustrial areaInput streamsImage Processing, Computer-AssistedHidden Markov modelsVisionVideo RecordingFusionintermethod comparisonhidden Markov modelecho state networkGenetic algorithmsWork-flowsComputer Simulationpriority journalcameraDatabases, FactualNeural Networks (Computer)ESNNovel methodsArtificial IntelligenceTime seriesHMMCalibrationActivity patternsAlgorithmsGenetic algorithmVision, OcularMulti-camerasPattern Recognition, AutomatedOnline classification of visual tasks for industrial workflow monitoringjournalArticle2011Δημοσίευση σε περιοδικόPublication in journalΤεχνητή νοημοσύνηArtificial Intelligenceisi - Computer Science, Artificial Intelligence covers resources that focus on research and techniques to create machines that attempt to efficiently reason, problem-solve, use knowledge representation, and perform analysis of contradictory or ambiguous information. This category includes resources on artificial intelligence technologies such as expert systems, fuzzy systems, natural language processing, speech recognition, pattern recognition, computer vision, decision-support systems, knowledge bases, and neural networks.Επιστημονικό άρθροScientific articleΕθνικό Μετσόβιο ΠολυτεχνείοGreek Aggregator OpenArchives.gr | National Documentation Centre (EKT)other