Level-Set Evolution with Region Competition: Automatic 3-D Segmentation of Brain Tumors 1 Sean Ho, 2 Elizabeth Bullitt, and 1;3 Guido Gerig 1 Department.

Slides:



Advertisements
Similar presentations
A Growing Trend Larger and more complex models are being produced to explain brain imaging data. Bigger and better computers allow more powerful models.
Advertisements

SPM5 Segmentation. A Growing Trend Larger and more complex models are being produced to explain brain imaging data. Bigger and better computers allow.
Ter Haar Romeny, FEV MIT AI Lab Automatic Polyp Detection.
VBM Voxel-based morphometry
MRI preprocessing and segmentation.
Gordon Wright & Marie de Guzman 15 December 2010 Co-registration & Spatial Normalisation.
Preprocessing: Coregistration and Spatial Normalisation Cassy Fiford and Demis Kia Methods for Dummies 2014 With thanks to Gabriel Ziegler.
These improvements are in the context of automatic segmentations which are among the best found in the literature, exceeding agreement between experts.
Jeroen Hermans, Frederik Maes, Dirk Vandermeulen, Paul Suetens
Assessing Early Brain Development in Neonates by Segmentation of High-Resolution 3T MRI 1,2 G Gerig, 2 M Prastawa, 3 W Lin, 1 John Gilmore Departments.
Hierarchical Statistical Modeling of Boundary Image Profiles Sean Ho Department of Computer Science University of North Carolina, Chapel Hill, NC, USA.
Sponsor: Prof. Sidney Spector Computational anatomy to assess growth pattern of early brain development in healthy and disease populations Guido Gerig.
University of North Carolina Comparison of Human and M-rep Kidneys Segmented from CT Images James Chen, Gregg Tracton, Manjari Rao, Sarang Joshi, Steve.
Medical Image Synthesis via Monte Carlo Simulation James Z. Chen, Stephen M. Pizer, Edward L. Chaney, Sarang Joshi Medical Image Display & Analysis Group,
September 27, / 18 Automatic Segmentation of Neonatal Brain MRI Marcel Prastawa 1, John Gilmore 2, Weili Lin 3, Guido Gerig 1,2 University.
Diffusion Tensor Imaging (DTI) is becoming a routine technique to study white matter properties and alterations of fiber integrity due to pathology. The.
Comp 775: Deformable models: snakes and active contours Marc Niethammer, Stephen Pizer Department of Computer Science University of North Carolina, Chapel.
Visualization Challenges: Specific Medical Needs and General Problems Elizabeth Bullitt University of North Carolina Chapel Hill
Cortical Surface Analysis and Automatic Parcellation of Human Brain Wen Li, Ph.D. student Advisor: Vincent A. Magnotta, Ph.D. Biomedical Engineering University.
Multimodal Visualization for neurosurgical planning CMPS 261 June 8 th 2010 Uliana Popov.
Preprocessing II: Between Subjects John Ashburner Wellcome Trust Centre for Neuroimaging, 12 Queen Square, London, UK.
Li Wang, Yaozong Gao, Feng Shi, Gang Li, Dinggang Shen
Texture-based Deformable Snake Segmentation of the Liver Aaron Mintz Daniela Stan Raicu, PhD Jacob Furst, PhD.
3D CT Image Data Visualize Whole lung tissues Using VTK 8 mm
Li Wang1, Feng Shi1, Gang Li1, Weili Lin1, John H
Voxel Based Morphometry
MNTP Trainee: Georgina Vinyes Junque, Chi Hun Kim Prof. James T. Becker Cyrus Raji, Leonid Teverovskiy, and Robert Tamburo.
MS Lesion Visualization Assisted Segmentation Daniel Biediger COSC 6397 – Scientific Visualization.
Vascular Attributes and Malignant Brain Tumors MICCAI November 2003 CONCLUSIONS References: [1] Aylward S, Bullitt E (2002) Initialization, noise, singularities.
 Process of partitioning an image into segments  Segments are called superpixels  Superpixels are made up several pixels that have similar properties.
Automatic Brain Segmentation in Rhesus Monkeys February 2006, SPIE Medical Imaging 2006 Funding provided by UNC Neurodevelopmental Disorders Research Center.
Anatomical Measures John Ashburner zSegmentation zMorphometry zSegmentation zMorphometry.
Quantitative analysis of radiologic images: Image segmentation and registration, statistical atlases Christos Davatzikos, Ph.D. Professor of Radiology.
DTU Medical Visionday May 27, 2009 Generative models for automated brain MRI segmentation Koen Van Leemput Athinoula A. Martinos Center for Biomedical.
NA-MIC National Alliance for Medical Image Computing Cortical Thickness Analysis Delphine Ribes (Internship UNC 2005/2006) Guido Gerig.
NA-MIC National Alliance for Medical Image Computing Cortical Thickness Analysis with Slicer Martin Styner UNC - Departments of Computer.
Image Segmentation and Seg3D Ross Whitaker SCI Institute, School of Computing University of Utah.
Constructing Image Graphs for Segmenting Lesions in Brain MRI May 29, 2007 Marcel Prastawa, Guido Gerig Department of Computer Science UNC Chapel Hill.
PVE for MRI Brain Tissue Classification Zeng Dong SLST, UESTC 6-9.
NA-MIC National Alliance for Medical Image Computing ABC: Atlas-Based Classification Marcel Prastawa and Guido Gerig Scientific Computing.
M. Pokric, P.A. Bromiley, N.A. Thacker, M.L.J. Scott, and A. Jackson University of Manchester Imaging Science and Biomedical Engineering Probabilistic.
National Alliance for Medical Image Computing Segmentation Foundations Easy Segmentation –Tissue/Air (except bone in MR) –Bone in CT.
NA-MIC National Alliance for Medical Image Computing Segmentation Core 1-3 Meeting, May , SLC, UT.
Integrating QDEC with Slicer3 Click to add subtitle.
Ventricular shape of monozygotic twins discordant for schizophrenia reflects vulnerability 2 M Styner, 1,2 G Gerig, 3 DW Jones, 3 DR Weinberger, 1 JA Lieberman.
NA-MIC National Alliance for Medical Image Computing Process-, Work-Flow in Medical Image Processing Guido Gerig
Incorporating Non-rigid Registration into Expectation Maximization Algorithm to Segment MR Images By K.M. Pohl, W.M. Wells, A. Guimond, K. Kasai, M.E.
NA-MIC National Alliance for Medical Image Computing Evaluating Brain Tissue Classifiers S. Bouix, M. Martin-Fernandez, L. Ungar, M.
Automatic pipeline for quantitative brain tissue segmentation and parcellation: Experience with a large longitudinal schizophrenia MRI study 1,2 G Gerig,
NA-MIC National Alliance for Medical Image Computing A longitudinal study of brain development in autism Heather Cody Hazlett, PhD Neurodevelopmental.
NA-MIC National Alliance for Medical Image Computing Slicer3 Tutorial Nonrigid Atlas Registration Dominik Meier, Ron Kikinis February.
NA-MIC National Alliance for Medical Image Computing Engineering a Segmentation Framework Marcel Prastawa.
MultiModality Registration Using Hilbert-Schmidt Estimators By: Srinivas Peddi Computer Integrated Surgery II April 6 th, 2001.
Detection of Anatomical Landmarks Bruno Jedynak Camille Izard Georgetown University Medical Center Friday October 6, 2006.
NA-MIC National Alliance for Medical Image Computing Analysis and Results of Brockton VA study: Controls vs Schizophrenics Personality Disorder Martin.
Asymmetric Bias in User Guided Segmentations of Subcortical Brain Structures May 2007, UNC/BRIC Radiology 2007 Funding provided by UNC Neurodevelopmental.
Validation and Evaluation of Algorithms
Random Forests For Multiple Sclerosis Lesion Segmentation
Pathology Spatial Analysis February 2017
Neuro Best Contrast Filter for Head CT
Registration of Pathological Images
MODULE II Semi-Automatic Segmentation
Detecting Gray Matter Maturation via Tensor-based Surface Morphometry
From buttons to code Eamonn Walsh & Domenica Bueti
Semi-Automatic Generation of Transfer Functions
Anatomical Measures John Ashburner
Fall 2018, COMP 562 Poster Session
MultiModality Registration using Hilbert-Schmidt Estimators
Segmentation Algorithm
Shengcong Chen, Changxing Ding, Minfeng Liu 2018
Presentation transcript:

Level-Set Evolution with Region Competition: Automatic 3-D Segmentation of Brain Tumors 1 Sean Ho, 2 Elizabeth Bullitt, and 1;3 Guido Gerig 1 Department of Computer Science, 2 Department of Surgery, 3 Department of Psychiatry University of North Carolina, Chapel Hill, NC, USA Supported by NIH-NCI R01 CA Partially supported by NIH-NCI P01 CA47982.

Tumor segmentation Focusing on meningiomas and glioblastomas Focusing on meningiomas and glioblastomas Glioblastomas have a ring enhancement that makes segmentation tough Glioblastomas have a ring enhancement that makes segmentation tough

Overview of the procedure 1. Multiparameter MR image data 2. Fuzzy voxel-based segmentation 3. Level-set snake driven by: 1. Region competition 2. Smoothness constraints Can use alone for enhancing tumors Can use alone for enhancing tumors Or as part of the tumor/tissue/vasculature segmentation Or as part of the tumor/tissue/vasculature segmentation

Multiparameter MR images T1GAD-T1 registered difference image T1GAD-T1 registered difference image T2 available but not used in this work T2 available but not used in this work = -

Probability map of enhancing tissue T1GAD-T1 registered difference image T1GAD-T1 registered difference image Mixture-model histogram fit: Mixture-model histogram fit: Gaussian for the background Gaussian for the background Gamma function for the contrast agent uptake Gamma function for the contrast agent uptake

Region competition snake Image force: modulate propagation by signed inside/outside force Image force: modulate propagation by signed inside/outside force Smoothness constraint: Smoothness constraint: Mean curvature flow Mean curvature flow Gaussian smoothing of the implicit function Gaussian smoothing of the implicit function

Enhancement => image force

Live demo

Results Very challenging segmentation problem, even for expert manual segmentation: Very challenging segmentation problem, even for expert manual segmentation: Complex tumor geometry Complex tumor geometry Complex greylevel appearance Complex greylevel appearance Nearby enhancing structures (e.g. vessels, bone) Nearby enhancing structures (e.g. vessels, bone) Some examples: Some examples:

Validation Compared against expert human rater Compared against expert human rater Validation with 2 nd human rater in progress Validation with 2 nd human rater in progress More tumor datasets on the way More tumor datasets on the way Dataset Volume Overlap Hausdorff (mm) In (mm) Out (mm) Average (mm) Tumor % Tumor % Tumor %

Integrating in the “ Big Picture ” Modify atlas with subject specific pathology Modify atlas with subject specific pathology Probability map of enhancing tissue Probability map of enhancing tissue Region-competition snake Region-competition snake Smoothness constraints Smoothness constraints EM tissue classification (previous talk): EM tissue classification (previous talk): Using spatial prior Using spatial prior Additional tumor and edema classes Additional tumor and edema classes Bias field inhomogeneity compensation Bias field inhomogeneity compensation Result: Combined tumor and tissue segmentation (gm, wm, csf, edema) Result: Combined tumor and tissue segmentation (gm, wm, csf, edema)

The “ Big Picture ”, cont. Tumor segmentation registered with segmentation of vasculature: Tumor segmentation registered with segmentation of vasculature: We also have MRA images We also have MRA images Vessel extraction software Vessel extraction software

Free software downloads midag.cs.unc.edu midag.cs.unc.edu SNAP (prototype): SNAP (prototype): 3D level-set evolution 3D level-set evolution Preprocessing pipeline and manual editing Preprocessing pipeline and manual editing VALMET (prototype) VALMET (prototype)