Jinsong Ouyang, PhD
Associate Professor of Radiology and Biomedical ImagingDownloadHi-Res Photo
Cards
Contact Info
About
Copy Link
Titles
Associate Professor of Radiology and Biomedical Imaging
Affiliated Faculty, Yale Institute for Global Health
Appointments
Radiology & Biomedical Imaging
Associate Professor on TermPrimary
Other Departments & Organizations
Education & Training
- PhD
- University of Colorado, Physics (1992)
Research
Copy Link
Overview
Originally trained in experimental nuclear and high-energy physics, I transitioned to medical imaging. I focus on algorithm development for PET and SPECT, including image reconstruction, parametric reconstruction, motion correction, partial-volume correction, attenuation correction, scatter correction, and lesion detection. In recent years, I have led multiple deep learning–based projects in medical imaging.
Public Health Interests
Metabolism; Modeling; Network Analysis; Cardiovascular Diseases; Cancer; Bioinformatics; Genetics, Genomics, Epigenetics; Statistical Computing; Stochastic Processes
ORCID
0000-0003-4245-170X
Research at a Glance
Yale Co-Authors
Frequent collaborators of Jinsong Ouyang's published research.
Publications Timeline
A big-picture view of Jinsong Ouyang's research output by year.
Georges El Fakhri, PhD, DABR
Thibault Marin, PhD
Xiaofeng Liu
Yanis Chemli, PhD
Chao Ma, PhD
Marc David Normandin, PhD
95Publications
1,578Citations
Publications
2026
Scan-wise generalized PET denoising with contrastive adversarial learning
Liu X, Marin T, Eslahi S, Tiss A, Chemli Y, Najmaoui Y, Jang S, Fakhri G, Ouyang J. Scan-wise generalized PET denoising with contrastive adversarial learning. Physics In Medicine And Biology 2026, 71: 105030. PMID: 42173146, PMCID: PMC13312193, DOI: 10.1088/1361-6560/ae7231.Peer-Reviewed Original ResearchAltmetricMeSH Keywords and ConceptsConceptsAdversarial frameworkContrastive lossDomain generalizationCross-entropyPeak signal-to-noise ratioSuperior denoising performanceContrastive learning schemeStructural similarity indexNoise realizationsAdversarial learningAdversarial trainingDenoising performanceSignal-to-noise ratioLearning schemeNegative pairsDeep learningAdversarial methodsPositive pairsStandard baselinesMutual informationDenoisingPerformance degradationSimilarity indexDistribution shiftsDomain distributionIndividualized treatment effect inference of head and neck cancer with multimodal data
Wei Y, Li Z, Woo J, Ouyang J, Fakhri G, Liu X. Individualized treatment effect inference of head and neck cancer with multimodal data. APSIPA Transactions On Signal And Information Processing 2026, 15: 3-20. PMID: 42125301, PMCID: PMC13160251, DOI: 10.1108/atsip-06-2025-0051.Peer-Reviewed Original ResearchConceptsHead and neck cancerAdversarial trainingMultimodal patient dataDeep learningHead and neck cancer casesMultimodal dataRobust estimates of causal effectsEnd-to-end deep learningMutual informationPatient-specific outcomesSelection biasEstimation of causal effectsAdaptive instance normalizationPatient dataNeck cancerMultimodal medical imagesEnd-to-endIndividual treatment effects estimationStatus featuresRetrospective observational dataInstance NormalizationTreatment effectsInformation fusionPatient characteristicsMedical imagesUnsupervised Adaptation from FDG to PSMA PET/CT for 3D Lesion Detection Under Label Shift
Liu X, Xia M, Chemli Y, Fakhri G, Liu C, Ouyang J. Unsupervised Adaptation from FDG to PSMA PET/CT for 3D Lesion Detection Under Label Shift. 2026, 00: 1-5. DOI: 10.1109/isbi61048.2026.11515918.Peer-Reviewed Original ResearchConceptsUnsupervised domain adaptationLabel shiftPseudo-labelsSupervised learningSelf-trainingPseudo-label selectionBox regressionDomain adaptationUnsupervised adaptationCovariate shiftConfidence thresholdLesion detectionLearningLabelingDetectionAnchor shapeUnsupervisedFROCHistogramAdaptationSize compositionPseudoAI‑driven multi-lesion detection in whole‑body FDG PET/CT
Liu X, Xia M, Chemli Y, Fakhri G, Liu C, Ouyang J. AI‑driven multi-lesion detection in whole‑body FDG PET/CT. Progress In Biomedical Optics And Imaging 2026, 13928: 7. DOI: 10.1117/12.3087729.Peer-Reviewed Original ResearchCitationsConceptsWhole-body FDG PET/CTFDG-PET/CTLesion detectionFDG-PET/CT studiesIntersection-over-unionOncologic PET/CTLesion detection networkDetection of lesionsDeep learning modelsCT informationDiagnostic accuracyPET-onlyTreatment planningLesion sizePET/CTObject detectorsEfficiency of radiologistsIoU thresholdLesionsDetection modelPublic datasetsDetectorDetection networkNumerous lesionsLocalization performanceExploring the limits of deep-learning‑based PET image denoising for lesion detectability
Bayerlein R, Xia M, Ouyang J, Chemli Y, Melnichuk D, Fakhri G, Nardo L, Liu C, Badawi R. Exploring the limits of deep-learning‑based PET image denoising for lesion detectability. Progress In Biomedical Optics And Imaging 2026, 13928: 5. DOI: 10.1117/12.3085222.Peer-Reviewed Original ResearchConceptsDenoised imageDL-based denoisersLesion contrastDetectability of low-contrast lesionsDL-basedInformation diffusion modelDeep learning denoisingPET image qualityImage qualityVisual image qualityArea under the ROC curveActivity concentration ratioLow-contrast lesionsOverall image appearanceNoisy imagesImage representationLearning denoisingDenoisingHigh-contrast featuresNoise levelLesion uptakeLesion-to-background ratioInput noise levelTOF-OSEMDetection taskIndividualized Treatment Effect Inference of HNC with Multimodal Data
Wei Y, Li Z, Woo J, Ouyang J, El Fakhri G, Liu X. Individualized Treatment Effect Inference of HNC with Multimodal Data. APSIPA Transactions On Signal And Information Processing 2026, 14 DOI: 10.1561/116.20250051.Peer-Reviewed Original Research
2025
Physics-Informed List-Mode Deep Image Prior Reconstruction with Motion Correction in 3D Brain PET
Chemli Y, Najmaoui Y, Normandin M, Fakhri G, Marin T, Ouyang J. Physics-Informed List-Mode Deep Image Prior Reconstruction with Motion Correction in 3D Brain PET. 2025 IEEE Nuclear Science Symposium (NSS), Medical Imaging Conference (MIC) And Room Temperature Semiconductor Detector Conference (RTSD) 2025, 1-2. DOI: 10.1109/nss/mic/rtsd57106.2025.11287851.Peer-Reviewed Original ResearchConceptsDeep Image PriorList-modeContrast recoveryMotion correctionList-mode eventsReconstructed activity distributionHoffman brain phantomNoise-resolution trade-offML-EM reconstructionLow-count dataResolution phantomBrain phantomBrain positron emission tomographySuper-resolution effectActivity distributionModel attenuationML-EMBack-projection operationsFine structureDown-sampled dataNegative log-likelihoodImage registrationUnsupervised regularizerBack-projectionImage priorsOn Hallucinations in Artificial Intelligence–Generated Content for Nuclear Medicine Imaging (the DREAM Report)
Xia M, Bayerlein R, Chemli Y, Liu X, Ouyang J, Lin M, Fakhri G, Badawi R, Li Q, Liu C. On Hallucinations in Artificial Intelligence–Generated Content for Nuclear Medicine Imaging (the DREAM Report). Journal Of Nuclear Medicine 2025, 67: 166-174. PMID: 41198241, PMCID: PMC12866389, DOI: 10.2967/jnumed.125.270653.Peer-Reviewed Original ResearchCitationsAltmetricYRT-PET: An Open-Source GPU-Accelerated Image Reconstruction Engine for Positron Emission Tomography
Najmaoui Y, Chemli Y, Toussaint M, Petibon Y, Marty B, Fontaine K, Gallezot J, Razdevsek G, Orehar M, Dhaynaut M, Guehl N, Dolenec R, Pestotnik R, Johnson K, Ouyang J, Normandin M, Tetrault M, Lecomte R, Fakhri G, Marin T. YRT-PET: An Open-Source GPU-Accelerated Image Reconstruction Engine for Positron Emission Tomography. IEEE Transactions On Radiation And Plasma Medical Sciences 2025, 10: 535-546. PMID: 41424471, PMCID: PMC12714321, DOI: 10.1109/trpms.2025.3619872.Peer-Reviewed Original ResearchThis study introduces YRT-PET, an open-source GPU-accelerated software for PET image reconstruction, demonstrating high flexibility, speed, and compatibility with existing tools for dynamic imaging and motion correction.Anatomically and metabolically informed diffusion for unified denoising and segmentation in low-count PET imaging
Xia M, Ko K, Wang D, Chen M, Liu Q, Xie H, Guo L, Ji W, Ouyang J, Bayerlein R, Spencer B, Li Q, Badawi R, Fakhri G, Liu C. Anatomically and metabolically informed diffusion for unified denoising and segmentation in low-count PET imaging. Medical Image Analysis 2025, 107: 103831. PMID: 41076965, PMCID: PMC12551811, DOI: 10.1016/j.media.2025.103831.Peer-Reviewed Original ResearchCitationsMeSH Keywords and ConceptsConceptsPET denoisingDenoising diffusion modelsSuperior performanceImage denoisingDenoised outputMulti-vendorDenoisingInformation diffusionImage informationSegmentation modelAblated versionsDiffusion strategySegmentation methodDice coefficientOrgan segmentationTest casesMulti-task functionRevision moduleTaskImagesAnalysis pipelineSegmentsClinical count levelsTotal lesion glycolysisDataset
News
Copy Link
News
Get In Touch
Copy Link