2023
Rethinking Semi-Supervised Medical Image Segmentation: A Variance-Reduction Perspective.
You C, Dai W, Min Y, Liu F, Clifton D, Zhou S, Staib L, Duncan J. Rethinking Semi-Supervised Medical Image Segmentation: A Variance-Reduction Perspective. Advances In Neural Information Processing Systems 2023, 36: 9984-10021. PMID: 38813114, PMCID: PMC11136570.Peer-Reviewed Original ResearchMedical image segmentationContrastive learningImage segmentationSemi-supervised medical image segmentationSemi-supervised contrastive learningSelf-supervised objectiveSemantic segmentation datasetsSemi-supervised methodGround-truth labelsQuality of visual representationSafety-critical tasksSegmentation datasetTail classesSegmentation taskLabel setsTruth labelsCL frameworkNegative examplesModel collapseVariance-reductionVariance-reduction techniquesVisual representationTaskLearningPairs of samples
2022
Simultaneous Segmentation and Motion Estimation of Left Ventricular Myocardium in 3D Echocardiography Using Multi-task Learning
Ta K, Ahn SS, Stendahl JC, Langdon J, Sinusas AJ, Duncan JS. Simultaneous Segmentation and Motion Estimation of Left Ventricular Myocardium in 3D Echocardiography Using Multi-task Learning. Lecture Notes In Computer Science 2022, 13131: 123-131. PMID: 35759335, PMCID: PMC9221412, DOI: 10.1007/978-3-030-93722-5_14.Peer-Reviewed Original ResearchMotion estimationMulti-task learning networkMedical image analysis literatureMulti-task learningSingle-task learningMotion estimation techniqueImage analysis literatureComputer visionDecoding branchesFeature encoderLearning frameworkLearning networkLatent featuresAccurate segmentationSimultaneous segmentationEstimate motionImage pairsTask learningRealistic motion patternsVolumetric segmentationSegmentationMotion patternsTaskUnique taskLearning
2020
Layer Embedding Analysis in Convolutional Neural Networks for Improved Probability Calibration and Classification
Zhang F, Dvornek N, Yang J, Chapiro J, Duncan J. Layer Embedding Analysis in Convolutional Neural Networks for Improved Probability Calibration and Classification. IEEE Transactions On Medical Imaging 2020, 39: 3331-3342. PMID: 32356739, PMCID: PMC7606489, DOI: 10.1109/tmi.2020.2990625.Peer-Reviewed Original ResearchConceptsConvolutional neural networkNeural networkClassification taskProbability calibrationTissue classification tasksImage representationBaseline methodsPublic datasetsModel performanceRandom forest modelNetworkBetter performanceForest modelDatasetClassificationTaskCT imagesImagesOriginal model outputMR imagesModel outputInstitutional datasetPerformanceEmbeddingOutputEstimating Reproducible Functional Networks Associated with Task Dynamics Using Unsupervised LSTMS
Dvornek NC, Ventola P, Duncan JS. Estimating Reproducible Functional Networks Associated with Task Dynamics Using Unsupervised LSTMS. 2011 IEEE International Symposium On Biomedical Imaging: From Nano To Macro 2020, 00: 1-4. PMID: 34422224, PMCID: PMC8375550, DOI: 10.1109/isbi45749.2020.9098377.Peer-Reviewed Original ResearchLong short-term memoryFunctional networksBiological motion perception taskTask activitiesMotion perception taskShort-term memoryLSTM modelPerception taskNeural correlatesTask paradigmFMRI activityTerm memoryRecurrent neural networkTask dynamicsTarget taskFunctional magnetic resonance imaging (fMRI) time series dataTaskUnsupervised mannerAdaptive Checkpoint Adjoint Method for Gradient Estimation in Neural ODE.
Zhuang J, Dvornek N, Li X, Tatikonda S, Papademetris X, Duncan J. Adaptive Checkpoint Adjoint Method for Gradient Estimation in Neural ODE. Proceedings Of Machine Learning Research 2020, 119: 11639-11649. PMID: 34308361, PMCID: PMC8299461.Peer-Reviewed Original ResearchNeural ordinary differential equationsComputation graphImage classification tasksClassification taskPyTorch implementationBenchmark tasksTraining timeAdaptive checkpointsNeural ODEAutomatic differentiationNaive methodTime series modelingRedundant componentsGradient estimation methodError rateGood accuracyPhysical knowledgeEmpirical performanceGraphGradient estimationTaskAccuracyODE solverSolverResNet
2019
Invertible Network for Classification and Biomarker Selection for ASD
Zhuang J, Dvornek NC, Li X, Ventola P, Duncan JS. Invertible Network for Classification and Biomarker Selection for ASD. Lecture Notes In Computer Science 2019, 11766: 700-708. PMID: 32274471, PMCID: PMC7144624, DOI: 10.1007/978-3-030-32248-9_78.Peer-Reviewed Original ResearchInvertible networksDeep learning methodsDeep learning modelsBlack-box natureLowest regression errorRegression tasksClassification taskLearning methodsLearning modelDecision boundariesModel decisionsImportant edgesLinear classifierConnectivity matrixASD classificationNetworkBlack-box representationBiomarker selectionRegression errorsData pointsImportance measuresTaskNovel methodClassificationClassifierDomain-Agnostic Learning with Anatomy-Consistent Embedding for Cross-Modality Liver Segmentation
Yang J, Dvornek NC, Zhang F, Zhuang J, Chapiro J, Lin M, Duncan JS. Domain-Agnostic Learning with Anatomy-Consistent Embedding for Cross-Modality Liver Segmentation. ICCV Workshops 2019, 00: 323-331. PMID: 34676308, PMCID: PMC8528125, DOI: 10.1109/iccvw.2019.00043.Peer-Reviewed Original ResearchDomain adaptationDisentangled representationsLiver segmentationTarget domainSource domainDeep learning modelsGenerative adversarial networkHuman interpretabilityLearning frameworkAdversarial networkDownstream tasksArt methodsSegmentation consistencyLearning modelAgnostic learningMeaningful representationCycleGANNew tasksAblation analysisDA taskDifferent modalitiesTaskSegmentationEmbeddingLearning
2018
Combining Phenotypic and Resting-State FMRI Data for Autism Classification with Recurrent Neural Networks
Dvornek NC, Ventola P, Duncan JS. Combining Phenotypic and Resting-State FMRI Data for Autism Classification with Recurrent Neural Networks. 2011 IEEE International Symposium On Biomedical Imaging: From Nano To Macro 2018, 2018: 725-728. PMID: 30288208, PMCID: PMC6166875, DOI: 10.1109/isbi.2018.8363676.Peer-Reviewed Original ResearchAutism spectrum disorderRecurrent neural networkNeural networkAutism Brain Imaging Data ExchangeSingle deep learning frameworkHeterogeneity of ASDFunctional magnetic resonance imagingDeep learning frameworkResting-state fMRI dataResting-state functional magnetic resonance imagingBetter classification accuracyAutism classificationSpectrum disorderData exchangeLearning frameworkFMRI dataClassification accuracyCross-validation frameworkChallenging taskStraightforward taskPrior workNetworkSuch dataRsfMRITask