2013
Numerical Surrogates for Human Observers in Myocardial Motion Evaluation From SPECT Images
Marin T, Kalayeh M, Parages F, Brankov J. Numerical Surrogates for Human Observers in Myocardial Motion Evaluation From SPECT Images. IEEE Transactions On Medical Imaging 2013, 33: 38-47. PMID: 23981533, PMCID: PMC4148467, DOI: 10.1109/tmi.2013.2279517.Peer-Reviewed Original ResearchMeSH KeywordsAlgorithmsBiomimeticsCoronary Artery DiseaseExpert SystemsHumansImage EnhancementImage Interpretation, Computer-AssistedMotionMyocardial Perfusion ImagingObserver VariationReproducibility of ResultsSensitivity and SpecificitySupport Vector MachineTomography, Emission-Computed, Single-PhotonConceptsHuman observer performanceRelevance vector machinePredicting human observer performanceDeformable mesh modelHuman observersHuman observer scoresImage-quality assessmentPerfusion-defect detectionMotion estimationModel observerMotion featuresVector machineMedical imagesLinear discriminantDefect detectionDetection taskMesh modelObserver performanceDiagnostic tasksNumerical surrogateTask-based approachApproximate surrogateTaskMotion evaluationPerformanceGeneralization Evaluation of Machine Learning Numerical Observers for Image Quality Assessment
Kalayeh M, Marin T, Brankov J. Generalization Evaluation of Machine Learning Numerical Observers for Image Quality Assessment. IEEE Transactions On Nuclear Science 2013, 60: 1609-1618. PMID: 25346545, PMCID: PMC4207371, DOI: 10.1109/tns.2013.2257183.Peer-Reviewed Original ResearchImage quality assessmentChannelized Hotelling observerHuman observer performanceVector machineMachine learningHuman observersTask-based evaluationEvaluation of medical imagesKernel-based regression modelsSupport vector machineInternal noise modelRelevance vector machineReduce model complexityMachine-learning frameworkBayesian machine-learning frameworkGeneralization performanceMedical imagesDetection taskQuality assessmentReconstructed imagesComputation timeHotelling observerNoise modelReconstruction methodModel complexity
2011
Channelized relevance vector machine as a numerical observer for cardiac perfusion defect detection task
Kalayeh M, Marin T, Pretorius P, Wernick M, Yang Y, Brankov J. Channelized relevance vector machine as a numerical observer for cardiac perfusion defect detection task. Proceedings Of SPIE--the International Society For Optical Engineering 2011, 7966: 79660e-79660e-7. DOI: 10.1117/12.878176.Peer-Reviewed Original Research
2010
Numerical observer for cardiac motion assessment using a linear discriminant
Marin T, Pretorius P, Yang Y, Wernick M, Brankov J. Numerical observer for cardiac motion assessment using a linear discriminant. 2010, 2352-2355. DOI: 10.1109/nssmic.2010.5874205.Peer-Reviewed Original ResearchDeformable mesh modelMesh modelDefect detection taskMotion estimation techniquesImage quality assessmentHuman visual systemHuman detection performanceHuman observersImage qualityMotion defectsHuman observer performanceChannelized Hotelling observerChannel modelMotion informationDetection performanceLinear discriminantDetection taskDiagnostic tasksHuman performanceProposed methodsSingle photon emission computed tomographyHotelling observerVisual systemEvaluate image qualityDiscriminant model