1. Computer Graphics Research at Virginia
David Luebke
Department of Computer Science

2. Outline Why I like graphics My current research VDS Gaze NPR
Scanning Monticello

3. Why Graphics? Graphics is cool Graphics is interesting
I like to see what I’m doing
I like to show people what I’m doing
Graphics is interesting
Involves simulation, algorithms, architecture…
Graphics is moving quickly
Intel vs. nVidia...
Graphics is fun
Graphics is important

4. Universal: Jurassic Park
Why Graphics?
Entertainment: Cinema
Universal: Jurassic Park
Pixar: Geri’s Game

5. Why Graphics?
Entertainment: Games
id: Quake II
Cyan: Riven

6. The Visible Human Project MIT: Image-Guided Surgery Project
Why Graphics?
Medical Visualization
The Visible Human Project
MIT: Image-Guided Surgery Project

7. Why Graphics?
Computer Aided Design (CAD)

8. Why Graphics?
Scientific Visualization

9. Outline Why I like graphics My current research
VDS: View-Dependent Simplification
Gaze
Art-Based Rendering
Scanning Monticello

10. The Problem: Massive Model Rendering
Problem: interactive visualization of very complex environments
Who cares?
Scientific visualization
Medical visualization
Computer Aided Design (CAD )

11. Massive Model Examples: Aerospace CAD
Cassini space probe model
415,000 polygons
Courtesy Jet Propulsion Laboratory

12. Massive Model Examples: Maritime CAD
Submarine Torpedo Room
700,000 polygons
Courtesy General Dynamics, Electric Boat Div.

13. Massive Model Examples: Structural CAD
Coal-fired power plant
13 million polygons
Courtesy ABB Engineering

14. Massive Model Examples: More Maritime CAD
Double Eagle container ship
82 million polygons
Courtesy Newport News Shipbuilding

15. The Holy Grail…
Deussen et al: Realistic Modeling of Plant Ecosystems

16. Level of Detail The problem: One solution:
Polygons rule the world (of interactive graphics)
Polygonal models are often too complex to render at interactive rates
One solution:
Level-of-detail methods simplify the polygonal geometry of small or distant objects

17. Level of Detail Traditional Approach
Create levels of detail (LODs) of objects:
249,924 polys
62,480 polys
7,809 polys
975 polys
Courtesy IBM

18. Level of Detail: Traditional Approach
Distant objects use coarser LODs:

19. Limitations of Traditional LOD
Most algorithms are:
Too fragile for messy CAD models
Too slow for large CAD models
Not suited for drastic simplification

20. Drastic Simplification: Large Objects
Courtesy IBM and ACOG

21. Drastic Simplification: Small Objects
Courtesy Electric Boat

22. Drastic Simplification: Preserving Topology
Rotor model:
21 holes
4736 faces
Courtesy Alpha_1 Project, University of Utah

23. Drastic Simplification: Preserving Topology
21 holes, 1006 faces
1 hole, 46 faces

24. The Problems With Drastic Simplification
For drastic simplification:
Large objects must be subdivided
Small objects must be combined
Topology must be simplified
Difficult or impossible with traditional LOD

25. A New Approach Dynamic: simplify objects on the fly
View-dependent: account for viewpoint
Global: simplify scenes, not objects
Automatic: simplify without user’s help

26. Dynamic Level of Detail
A relatively recent departure from the traditional static approach:
Static LOD: create individual LODs in a preprocess
Dynamic LOD: create data structure from which a desired level of detail can be extracted at run time.

27. View-Dependent LOD: Examples
Show nearby portions of object at higher resolution than distant portions
View from eyepoint
Birds-eye view

28. View-Dependent LOD: Examples
Show silhouette regions of object at higher resolution than interior regions

29. View-Dependent LOD: Examples
Show more detail where the user is looking than in their peripheral vision:
11,726 triangles

30. View-Dependent LOD: Examples
Show more detail where the user is looking than in their peripheral vision:
34,321 triangles

31. Dynamic LOD: How Does It Work?
So…dynamic LOD is great, but how do we implement it?
Basically, with one big data structure: the vertex tree
Represents the entire model
Hierarchy of all vertices in model
Queried each frame for updated scene

32. The Vertex Tree Each vertex tree node contains:
A subset of model vertices
A representative vertex or repvert
Folding a node collapses its vertices to the repvert
Unfolding a node splits the repvert back into vertices

33. Vertex Tree Example Triangles in active list Vertex tree 8 7 R 2 I II10 6 9 3 10 A B C 3 1 1 2 7 4 5 6 8 9 4 5
Triangles in active list
Vertex tree

34. Vertex Tree Example Triangles in active list Vertex tree 8 7 R A 2 III 10 6 9 3 10 A B C 3 1 1 2 7 4 5 6 8 9 4 5
Triangles in active list
Vertex tree

35. Vertex Tree Example Triangles in active list Vertex tree 8 R A I II 106 9 3 10 A B C 3 1 2 7 4 5 6 8 9 4 5
Triangles in active list
Vertex tree

36. Vertex Tree Example Triangles in active list Vertex tree 8 R A I II 106 9 3 10 A B C 3 B 1 2 7 4 5 6 8 9 4 5
Triangles in active list
Vertex tree

37. Vertex Tree Example Triangles in active list Vertex tree 8 R A I II 109 3 10 A B C 3 B 1 2 7 4 5 6 8 9
Triangles in active list
Vertex tree

38. Vertex Tree Example Triangles in active list Vertex tree 8 R A C I II10 9 3 10 A B C 3 B 1 2 7 4 5 6 8 9
Triangles in active list
Vertex tree

39. Vertex Tree Example Triangles in active list Vertex tree R A C I II 103 10 A B C 3 B 1 2 7 4 5 6 8 9
Triangles in active list
Vertex tree

40. Vertex Tree Example Triangles in active list Vertex tree R A C II I II10 3 10 A B C 3 B 1 2 7 4 5 6 8 9
Triangles in active list
Vertex tree

41. Vertex Tree Example Triangles in active list Vertex tree R A II I II10 10 A B C 3 B 1 2 7 4 5 6 8 9
Triangles in active list
Vertex tree

42. Vertex Tree Example Triangles in active list Vertex tree R A I II I II10 10 A B C 3 B 1 2 7 4 5 6 8 9
Triangles in active list
Vertex tree

43. Vertex Tree Example Triangles in active list Vertex tree R I II I II10 A B C 3 B 1 2 7 4 5 6 8 9
Triangles in active list
Vertex tree

44. Vertex Tree Example Triangles in active list Vertex tree R I II I II R10 A B C 3 B 1 2 7 4 5 6 8 9
Triangles in active list
Vertex tree

45. Vertex Tree Example Triangles in active list Vertex tree R I II R 10 AB C 3 1 2 7 4 5 6 8 9
Triangles in active list
Vertex tree

46. The Vertex Tree: Tris and Subtris8 7 9 8 10 5 4 6 A 3
Fold Node A 2 10 9 6 3 1
Unfold Node A 4 5
Node->Tris: triangles that change shape upon folding
Node->Subtris: triangles that disappear completely

47. The Vertex Tree: Tris and Subtris
Each node’s tris and subtris can be computed offline to be accessed very quickly at run time
This is the key observation behind dynamic simplification

48. Outline Why I like graphics My current research VDS
Gaze: gaze-directed rendering
NPR
Scanning Monticello

49. Gaze-Directed Rendering
View-dependent rendering opens up a new door: gaze-directed rendering
Track the user’s eye gaze
Give more detail to region of interest, less elsewhere

50. Gaze-Directed Rendering: ERICA
Problem 1: how to do eye tracking?
Answer ERICA project (Prof. Hutchinson)
Main uses: disabled access, gaze analysis
ERICA pros:
non-intrusive eye tracking
60 Hz, ~1 cm accuracy
ERICA cons:
No windows (working on it)
Head must stay very still (working on it)
Currently a bit bulky (working on it)

51. Gaze-Directed Rendering: Perceptual Science
Problem 2: how to degrade detail?
Want a rigorous solution, not a hack
Goal 1: Imperceptible LOD
Idea: use findings from perceptual physiology
Map contrast gratings to nodes in tree, only fold nodes when gratings are imperceptible
Goal 2: principled approach to polygon-budget rendering
This is ongoing work in the truest sense
Paper underway
Future: better perceptual mode, other rendering modalities

52. Outline Why I like graphics My current research VDS Gaze
NPR: Non-Photorealistic Rendering
Scanning Monticello

53. Non-Photorealistic Rendering
Fancy name, simple idea:
Make computer graphics that don’t look like computer graphics

54. Non-Photorealistic Rendering
Fancy name, simple idea:
Make computer graphics that don’t look like computer graphics

55. Non-Photorealistic Rendering
Fancy name, simple idea:
Make computer graphics that don’t look like computer graphics

56. Bunny: Traditional CG Rendering
NPRlib
NPRlib: flexible callback-driven NP rendering
Bunny: Traditional CG Rendering

57. Non-Photorealistic Rendering
NPRlib: flexible callback-driven NP rendering
Bunny: Pencil-Sketch Rendering

58. Non-Photorealistic Rendering
NPRlib: flexible callback-driven NP rendering
Bunny: Charcoal Smudge Rendering

59. Non-Photorealistic Rendering
NPRlib: flexible callback-driven NP rendering
Bunny: Two-Tone Rendering

60. Non-Photorealistic Rendering
NPRlib: flexible callback-driven NP rendering
Molecule: Traditional CG Rendering

61. Non-Photorealistic Rendering
NPRlib: flexible callback-driven NP rendering
Molecule: Painterly Rendering

62. Non-Photorealistic Rendering
NPRlib: flexible callback-driven NP rendering
Molecule: Cartoon Rendering

63. NPRlib Senior thesis, Derek Cornish: Ongoing research, Andrea Rowan
Turns out that VDSlib is an excellent foundation on which to build NPRlib
Particles to guide strokes  vertices
Regulate screen-space density  view-dependent simplification
Ongoing research, Andrea Rowan
We have a start, need:
Better architecture
Better examples
Users

64. Outline Why I like graphics My current research VDS Gaze
NPR: Non-Photorealistic Rendering
Scanning Monticello

65. Scanning Monticello Fairly new technology: scanning the world
My connection: telepresence project at UNC

66. Scanning Monticello Want a flagship project to showcase this
Idea: scan Thomas Jefferson’s Monticello
Historic preservation
Virtual tours
Archeological and architectural research, documentation, and dissemination
Great driving problem for scanning & rendering research
Just finished first pilot project
Show some data...

67. Outline Why I like graphics My current research MMR Gaze
NPR: Non-Photorealistic Rendering
Scanning Monticello

68. Graphics Resources Several SGI O2 workstations 2 SGI Octanes
Midrange graphics hardware
SGI InfiniteReality2
2 x 225 MHz R10K, 1 Gb, 4 Mb cache
High-end graphics hardware: 13 million triangles/sec, 64 Mb texture memory
Hot new PC platforms
High-end card built on nVidia’s best chipset
Stereo glasses, digital video card, miniDV stuff
Quad Xeon on loan

69. Breakout Hackfest: Wednesdays @ 5:30 PM Graphics Lunch: Fridays @ noon
Let’s schedule a breakout and do the pizza thing next week…