You're about to create your best presentation ever

Martinos Center Presentation Template

Create your presentation by reusing one of our great community templates.

4-D PET-MR with volumetric navigators and compressive sensing.

Transcript: Keith Heberlein MARTINOS CENTER, SIEMENS vNAV 275 ms .. Reg. 80 ms EXAMPLE PET RECONSTRUCTION TI gap MOTION-FREE PET MR MARTINOS CENTER, SIEMENS -OCCIPUT.IO- Dylan Tisdall MARTINOS CENTER MARTINOS CENTER MARTINOS CENTER, AALTO UNIV., DTU SLOW DRIFT COUNTS alternated direction method of multipliers (ADMM) SPORADIC EVENTS - DIRECT COMPRESSION TI gap PET RAY-TRACER iLang PET RAY-TRACER Stefano Pedemonte, Ciprian Catana, Koen Van Leemput MOTION SENSING MARTINOS CENTER MARTINOS CENTER automated algorithm synthesis LATEX PILLOW (3 cm thick) ACKNOWLEDGEMENTS Ciprian Catana MICCAI 2014 -VOLUMETRIC NAVIGATORS- Reg. 80 ms TI gap SOFTWARE vNAV TR gap Web-based computing engine for medical imaging -SPATIO-TEMPORAL TOMOGRAPHY- Himanshu Bhat FREE MOTION INFERENCE LANGUAGE MARTINOS CENTER MPRAGE readout TR gap DIRECT 4D 4-D Generative Model for PET/MRI Reconstruction, MICCAI 2011 TR Andre Van Der Kouwe INFERENCE LANGUAGE - HEALTHY VOLUNTEER automated synthesis -RECONSTRUCTION AS REASONING- - vNAV vNAV 275 ms Shirley Hsu vNAV 275 ms ACTIVITY inv. pulse - UNIFIED GEOMETRY ALGORITHM SYNTHESYS *at the core of the Insight Reconstruction Toolkit (RTK) -PROJECTION REVISITED- PERFORMANCE of NiftyRec*: PHANTOM - CYCLIC MOTION inv. pulse - 4-D SINOGRAM - Martinos Center for Biomedical Imaging, MGH - Harvard University - Aalto Univerisity, School of Science and Technology - Technical University of Denmark (DTU) Reg. 80 ms non-negativity, sparsity 4-D PET-MR with Volumetric Navigators and Compressed Sensing Koen Van Leemput MODEL BASED MOTION MPRAGE readout Computational Methods for Molecular Imaging fast convergence Explicit deformation model Grae Arabasz inv. pulse iLang MPRAGE readout vNAV Likelihood Gradient Activity Gradient Motion projection projection projection back-projection Likelihood: projection back-projection Gradient: activity photon counts MICCAI BAMBI workshop Thursday 2 pm MOTION EVENT DISCRIMINATION

An Inference Language for Imaging.

Transcript: Keith Heberlein MARTINOS CENTER, SIEMENS vNAV 275 ms .. Reg. 80 ms TI gap SAMPLING - - MARTINOS CENTER, SIEMENS HEALTHY VOLUNTEER MARTINOS CENTER INFERENCE LANGUAGE MARTINOS CENTER Dylan Tisdall MARTINOS CENTER, AALTO, DTU SLOW DRIFT - LATEX PILLOW (3 cm thick) COUNTS iLang - inference language - transmission - cone beam - parallel beam - emission - parallel beam - sparse projections 2D The graph within the graph RIEMANNIAN MANIFOLD TOMOGRAPHY MICCAI 2014 - BAMBI TI gap geometric primitives The algorithm arises from a model and from an inference query. MARTINOS CENTER MARTINOS CENTER iLang - inference language ACKNOWLEDGEMENTS GPU Ciprian Catana MIT MRI Reg. 80 ms TI gap 4-D PET-MR LANGUAGE PRIMITIVES iLang - inference language TR gap MARTINOS CENTER MASS-MATRIX = TOEPLIZ APPROX. OF EMPIRICAL FISHER INFORMATION - Himanshu Bhat - generative .. - model based .. - unified .. - joint .. - Bayesian MPRAGE readout TR gap iLang - inference language Registration Andre Van Der Kouwe X-Ray CT 1 - MAXIMUM PROBABILITY - ALTERNATING DIRECTION METHOD OF MULTIPLIERS - PRECONDITIONED GRADIENT DESCENT - LBFGS (QUASI NEWTON) - NEWTON - Affine - B-splines - 3-linear resampling vNAV 275 ms Shirley Hsu - Registration vNAV 275 ms - Ray-tracing: - Transformation: - Differentiation: - Stencil operations - FFT - IFFT - GEOMETRIC PRIMITIVES ACTIVITY inv. pulse - Martinos Center for Biomedical Imaging, MGH - Harvard University - Aalto Univerisity, School of Science and Technology - Technical University of Denmark (DTU) 4-D PET Paulina Golland's Journal Club PET RECONSTRUCTION REGISTRATION Probabilistic reasoning. RIPETITIVE MOTION PERFORMANCE inv. pulse 4-D SINOGRAM TWO TYPES OF INFERENCE QUERIES - PET Reg. 80 ms INFERENCE ENGINE 2 - POSTERIOR SAMPLING HAMILTONIAN MARKOV CHAIN MONTE CARLO MESSAGE PASSING Koen Van Leemput ray-tracing: NiftyRec An Inference Language for Imaging - - UNIFIED GEOMETRY - COMPRESSED 4-D SINOGRAM 4-D nav-free PET-MR MPRAGE readout a data structure? MOTION PET ray tracer Grae Arabasz Stefano Pedemonte, Ciprian Catana, Koen Van Leemput inv. pulse MPRAGE readout - finite differences - gradient propagation - PET RAY-TRACER MOTION EVENT DISCRIMINATION

Now you can make any subject more engaging and memorable