Introduction Image Analysis & Computer VisionSlide Number 2CS/BIOEN 6640 F2010CS/BIOEN 6640 F2010GoalsImage SensorsDigital ImageDigital ImageDigital ImagesDigital ImagesEdges: Sudden change of intensity LSegmentation of structuresSlide Number 13Slide Number 14Slide Number 16Concept of level-set evolutionSegmentation toolDemo itkSNAP toolSlide Number 20Results Brain Tumor SegmentationVentricle Segmentation by 3D Snakes: UNC SNAP ToolUse of deformable models in Vision IUse of deformable models in Vision IISlide Number 25Slide Number 26Blurring is diffusionBlurring of imagesLinear DiffusionWe want noise reduction while keeping structure boundariesNonlinear DiffusionSlide Number 32Shape from silhouettesMotivation: MoviesWhat is shape from silhouette?Visual hull as voxel gridSlide Number 37Metric Cameras and Visual-Hull Reconstruction from 4 viewsSlide Number 39Using probabilistic shape modelsNatural Shape VariabilityNotion of Shape SpaceFirst Eigenmode of Deformation (CC)Slide Number 57Automatic deformable model based 2D segmentationImage ProcessingWhy Image Analysis?Next Lecture Thu Aug 25Introduction Image Analysis & Computer VisionGuido GerigCS/BIOEN 6640 FALL 2010August 23, 2010Scientific Computing and Imaging Institute, University of UtahCourses and Seminars related to Research in Image AnalysisNEW in 2010: SoC Image Analysis Track (Director Tom Fletcher) (click)Fall 2010:• Image Processing CS 6640/ BIOEN 6640Spring 2011:• 3D Computer Vision CS 6320• Advanced Image Processing CS 6640• Mathematics of Imaging BIOEN 6500Fall 2011:• Image Processing Basics CS 4961• Image Processing CS 6640On demand:• Special Topics Courses: Non-Euclidean Geometry, Non-Param. Stats, ..Seminars:• Seminar Imaging “ImageLunch” CS 7938: weekly Mondays 12 to 1.15, WEB 3670Scientific Computing and Imaging Institute, University of UtahCS/BIOEN 6640 F2010For class: • 1) Go to the web-site page: http://www.sci.utah.edu/~gerig/CS6640-F2010/CS6640-F2010.html • 2) Look over the instructions and syllabus• 3) Follow the link to "mailing lists" and join the cs6640 mailing lists as in the instructions. Remind them to use a mail address that they actually read (COMING SOON)• 4) Look at the final and midterm exam dates and mark those on your calendar • 5) Purchase the book• 6) Do the first 2 reading assignments.Scientific Computing and Imaging Institute, University of UtahCS/BIOEN 6640 F2010For class: • We will use the uxxxxxxxx email address for communication, please forward the u-email to your personal email if you use another account.• The web-site provides downloads for additional materials and handouts. • The syllabus is not completely rigid and fixed, and some topics will develop as the class continues. • We will primarily use MATLAB (no extensions and additional libraries) for the projects. You can use CADE lab licenses or purchase a personal student license. C++ is an option (see web-page).• Etc.Scientific Computing and Imaging Institute, University of UtahGoals• to tell you what you can do with digital images• to show you that doing research in computer vision and image analysis is fun and exciting• to demonstrate that image processing is based on strong mathematical principles, applied to digital images via numerical schemes•to show you that you can solve typical image processing tasks on your ownScientific Computing and Imaging Institute, University of UtahImage SensorsScientific Computing and Imaging Institute, University of UtahDigital ImageScientific Computing and Imaging Institute, University of UtahDigital ImageEach cell has number, either a scalar (black and white) or a vector (color).Discrete representation of continuous world (sampling with aperture).Scientific Computing and Imaging Institute, University of UtahDigital ImagesScientific Computing and Imaging Institute, University of UtahDigital ImagesScientific Computing and Imaging Institute, University of UtahEdges: Sudden change of intensity LScientific Computing and Imaging Institute, University of UtahSegmentation of structures• User painting/drawing on 2D images (“photoshop”)• Tedious, time consuming, limited precision• Demonstrate ToolScientific Computing and Imaging Institute, University of UtahDeformable Models: SNAKESGeodesic Snake formulated as PDE[ ]→=∂∂Nttxcα),(Curve evolves over timeNormal direction to curveSpeedScientific Computing and Imaging Institute, University of UtahDeformable Models: SNAKESGeodesic Snake:[ ]→=∂∂Nttxcκ),(Curve evolves over timeNormal direction to curveCurvature (convex, concave)Mathematical solution is circleScientific Computing and Imaging Institute, University of UtahDeformable Models: SNAKESGeodesic Snake:Plus: add a term that stops at boundaries[ ]Ntcακ+=∂∂Scientific Computing and Imaging Institute, University of UtahConcept of level-set evolutionImplementation: Curve C embedded as zero-level of higher order function ϕScientific Computing and Imaging Institute, University of UtahSegmentation tool• User painting slice by slice (“photoshop”)• Tedious, time consuming, limited reproducibility• Painting in 2D intuitive, but what about 3D?ventriclesSo far: Slice-by-slice contouringScientific Computing and Imaging Institute, University of UtahDemo itkSNAP toolScientific Computing and Imaging Institute, University of Utah3D Geodesic SnakeChallenges:• efficient, stable 2D/3D implementation (implicit, fast marching,..)• appropriate image match function to stop propagation( ) ( ) ( )ϕϕαϕϕϕϕϕ21)()( ∇⋅+∇+∇⋅∇+∇∇∇⋅∇=∂∂∇+scgMCFrsrrrgcggggtScientific Computing and Imaging Institute, University of UtahResults Brain Tumor SegmentationT2EdemaTumorT13DPrastawa et al., Media 2004Type 1Type 2Type 3Scientific Computing and Imaging Institute, University of UtahVentricle Segmentation by 3D Snakes: UNC SNAP ToolInitia-lization by bubblesFinal Segmen-tation (10 seconds)2D axial MRI (3T MPrage)3D surface rendering3D surface rendering2D axial MRI (3T MPrage)Reliability: 0.99Efficiency: 2 MinDownload: http://www.ia.unc.edu/devScientific Computing and Imaging Institute, University of UtahUse of deformable models in Vision IScientific Computing and Imaging Institute, University of UtahUse of deformable models in Vision IIScientific Computing and Imaging Institute, University of UtahScientific Computing and Imaging Institute, University of UtahImage NoiseScientific Computing and