Overview Credits News Lecture 14: Scientific Visualization What - PowerPoint PPT Presentation

Overview Credits News Lecture 14: Scientific Visualization • What is SciVis? • almost unchanged from lecture by • Reminder: no class next week Melanie Tory (University of Victoria) – I'm at InfoVis/Vis in Baltimore • Data & Applications Information Visualization – who in turn used resources from CPSC 533C, Fall 2006 • Iso-surfaces – Torsten Möller (Simon Fraser University) • Direct Volume Rendering – Raghu Machiraju (Ohio State University) Tamara Munzner – Klaus Mueller (SUNY Stony Brook) • Vector Visualization UBC Computer Science • Challenges 26 Oct 2006 Difference between SciVis and Difference between SciVis and Overview Medical Scanning InfoVis InfoVis • What is SciVis? • MRI, CT, SPECT, PET, ultrasound • Card, Mackinlay, & Shneiderman: Parallel Direct Volume • Data & Applications Coordinates – SciVis: Scientific, physically based Rendering [Hauser et al ., – InfoVis: Abstract Vis 2000] [Fua et al. , Vis • Iso-surfaces Isosurfaces 1999] • Munzner: Glyphs Scatter Plots Line Integral • Direct Volume Rendering – SciVis: Spatial layout given [http://www.axon.com Convolution / – InfoVis: Spatial layout chosen Node-link gn_Acuity.html] • Vector Visualization [Cabral & Leedom, SIGGRAPH 1993] Diagrams Streamlines • Tory & Möller: • Challenges [Lamping et al. , CHI 1995] – SciVis: Spatial layout given + Continuous [Verma et al ., Vis 2000] SciVis InfoVis – InfoVis: Spatial layout chosen + Discrete – Everything else -- ? Medical Scanning - Medical Scanning - Biological Scanning Industrial Scanning Applications Applications • Medical education for anatomy, surgery, etc. • Surgical simulation for treatment planning • Scanners: Biological scanners, electronic microscopes, • Planning (e.g., log scanning) confocal microscopes • Illustration of medical procedures to the patient • Tele-medicine • Quality control • Inter-operative visualization in brain surgery, • Apps – physiology, paleontology, microscopic analysis… • Security (e.g. airport scanners) biopsies, etc. Scientific Computation - Scientific Computation - Apps Overview Isosurfaces - Examples Domain • Mathematical analysis • Flow Visualization • What is SciVis? Isolines Isosurfaces • ODE/PDE (ordinary and partial • Data & Applications differential equations) • Iso-surfaces • Finite element analysis (FE) • Direct Volume Rendering • Supercomputer simulations • Vector Visualization • Challenges

Isosurface Extraction MC 1: Create a Cube MC 2: Classify Each Voxel MC 3: Build An Index 0 1 1 3 2 • by contouring • Consider a Cube defined by eight data values: • Classify each voxel according to whether it lies • Use the binary labeling of each voxel to create an index outside the surface (value > iso-surface value) – closed contours inside the surface (value <= iso-surface value) – continuous 1 3 6 6 3 – determined by iso-value (i,j+1,k+1) (i+1,j+1,k+1) 10 10 v8 v7 • several methods 11110100 3 7 9 7 3 inside =1 Iso=9 (i,j,k+1) – marching cubes is most (i+1,j,k+1) 5 v4 5 v3 outside=0 common 2 7 8 6 2 10 v5 8 v6 00110000 Iso=7 (i,j+1,k) (i+1,j+1,k) 1 2 3 4 3 8 8 v1 v2 Index: =inside Iso-value = 5 =outside v1 v2 v3 v4 v5 v6 v7 v8 (i,j,k) (i+1,j,k) MC 4: Lookup Edge List MC 4: Example MC 5: Interp. Triangle Vertex MC 6: Compute Normals • For a given index, access an array storing a list of edges • Index = 00000001 • For each triangle edge, find the vertex location along the edge using linear • Calculate the normal at each cube vertex interpolation of the voxel values • triangle 1 = a, b, c c a G v v = � i i+1 x x i 1 , j , k i 1 , j , k + � G v v = � y i , j + 1 , k i , j � 1 , k b =10 G v v = � =0 z i , j , k + 1 i , j , k � 1 T v [] i • Use linear interpolation to compute the polygon T=8 � � � T=5 x i = + � � vertex normal � ] [] � [ • all 256 cases can be derived from 15 base cases v i 1 v i + � � � Direct Volume Rendering MC 7: Render! Overview Rendering Pipeline (RP) Examples • What is SciVis? Classify • Data & Applications • Iso-surfaces • Direct Volume Rendering • Vector Visualization • Challenges Classification Transfer Functions (TF’s) TF’s Transfer Function Challenges RGB � • original data set has application specific • Simple (usual) case: Map data • Setting transfer functions is difficult, unintuitive, • Better interfaces: values (temperature, velocity, proton density, value f to color and opacity and slow – Make space of TFs less confusing etc.) � f – Remove excess “flexibility” � • assign these to color/opacity values to make – Provide guidance sense of data RGB( f ) � ( f ) f • Automatic / semi-automatic transfer function generation • achieved through transfer functions f – Typically highlight boundaries � � Shading, Compositing… f f Gordon Kindlmann Gordon Kindlmann Gordon Kindlmann Human Tooth CT

Rendering Pipeline (RP) Light Effects Rendering Pipeline (RP) Interpolation • Usually only considering 2D 1D Light reflected part Classify reflected Classify • Given: specular • Given: Light Shade absorbed Shade ambient diffuse transmitted Interpolate • Needed: Light=refl.+absorbed+trans. Light=ambient+diffuse+specular • Needed: I k I k I k I = + + a a d d s s Interpolation Rendering Pipeline (RP) Ray Traversal Schemes Ray Traversal - First Intensity Intensity • Very important; regardless of algorithm Max Classify • Expensive => done very often for one image • Requirements for good reconstruction Average – performance Shade – stability of the numerical algorithm First – accuracy Linear Accumulate Nearest Interpolate First Depth neighbor • First : extracts iso-surfaces (again!) done by Tuy&Tuy ’84 Composite Depth Ray Traversal - Average Ray Traversal - MIP Ray Traversal - Accumulate Volumetric Ray Integration Intensity Intensity Intensity Max Average color opacity Accumulate 1.0 Depth Depth Depth • Max : Maximum Intensity Projection • Average : produces basically an X-ray picture • Accumulate : make transparent layers visible! used for Magnetic Resonance Angiogram Levoy ‘88 object (color, opacity) Overview Flow Visualization Traditional Flow Experiments Techniques • What is SciVis? • Traditionally – Experimental Flow Vis • Data & Applications • Now – Computational Simulation • Iso-surfaces • Typical Applications: • Direct Volume Rendering Glyphs (arrows) – Study physics of fluid flow Contours • Vector Visualization – Design aerodynamic objects • Challenges Jean M. Favre Streamlines

Techniques Techniques - Stream-ribbon Techniques - Stream-tube Mappings - Flow Volumes • Trace one streamline and a constant size • Generate a stream-line and widen it to a tube • Instead of tracing a line - trace a small vector with it • Width can encode another variable polyhedron • Allows you to see places where flow twists Overview Challenges - Accuracy Challenges - Accuracy LIC (Line Integral Convolution) • Need metrics -> perceptual metric • Integrate noise texture along a streamline • What is SciVis? • Deal with unreliable data (noise, • Data & Applications ultrasound) • Iso-surfaces • Direct Volume Rendering • Vector Visualization • Challenges (a) Original (c) Edge-Distorted (b) Bias-Added H.W. Shen Challenges - Accuracy Challenges - Speed/Size Challenges - HCI Challenges - HCI • “Augmented” reality • Irregular data sets • Efficient algorithms • Need better Structured Grids: • Explore novel I/O devices • Hardware developments (VolumePro) interfaces • Utilize current hardware (nVidia, ATI) • Which method • Compression schemes is best? regular uniform rectilinear curvilinear • Terabyte data sets Unstructured Grids: regular irregular hybrid curved