DOC PREVIEW
Princeton COS 426 - Traditional Computer Graphics

This preview shows page 1-2-3 out of 9 pages.

Save
View full document
View full document
Premium Document
Do you want full access? Go Premium and unlock all 9 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 9 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 9 pages.
Access to all documents
Download any document
Ad free experience
Premium Document
Do you want full access? Go Premium and unlock all 9 pages.
Access to all documents
Download any document
Ad free experience

Unformatted text preview:

1Image-Based Modeling and RenderingThomas FunkhouserPrinceton UniversityCOS 426 Guest LectureSpring 2003Traditional Computer Graphics(McMillan)When Does This Pipeline Work?How would you model and render this scene?(Jensen)When Does This Pipeline Work?How about this one?(Jensen)When Does This Pipeline Work?.. and this one?(Kim)When Does This Pipeline Work?What about this one?(Jensen)2When Does This Pipeline Work?How about this one?(Louvre)When Does This Pipeline Work?(McMillan)When doesn’t this pipeline work?Geometric Modeling(Levo y)It is hard to create 3D meshes for complex objects ...Reflectance Modeling(Square)It is hard to acquire good BRDFs for complex surfaces ...Light Transport Simulation(RenderPark)It is hard to compute all light paths for complex illumination ...What Else Can We Do?(McMillan)3Image-Based Rendering (IBR)• Model scene as set of reference images• Render novel views by resampling pixelsNovel viewReference imagesIBR Rendering Pipeline(McMillan)IBRModelGeometricModelImage-Based Representations• Plenoptic function (7D):o Describes the radiance traveling along a ray - to/from any point (x, y, z),- in any direction (φ, θ),- at any frequency (λ),- at any time (t)F(x, y, z, φ, θ, λ, t)(x,y,z)(φ,θ)Image-Based RepresentationsF(x, y, z, φ, θ, λ, t)7D6D5D4D3D2DIdealConsider only 3 frequencies (RGB)Consider rays through scene at one time instantConsider only lines through sceneConsider one dimension fewer directions/positionsConsider only directions from viewpointImage-Based RepresentationsF(x, y, z, φ, θ, λ, t)7D6D5D4D3D2DIdealConsider only 3 frequencies (RGB)Consider rays through scene at one time instantConsider only lines through sceneConsider one dimension fewer directions/positionsConsider only directions from viewpointImage PanoramasCylindrical Panorama(QuicktimeVR)4Kang 99Image Panoramas Image PanoramasVirtual tours of Princeton(Brow n)Image-Based RepresentationsF(x, y, z, φ, θ, λ, t)7D6D5D4D3D2DIdealConsider only 3 frequencies (RGB)Consider rays through scene at one time instantConsider only lines through sceneConsider one dimension fewer directions/positionsConsider only positions on surfacesTexture Mapping• Map photographs onto surfacesGeometry +Texture MapsPhotographsTexture Mapping(D ebevec)Image-Based RepresentationsF(x, y, z, φ, θ, λ, t)7D6D5D4D3D2DIdealConsider only 3 frequencies (RGB)Consider rays through scene at one time instantConsider only lines through sceneConsider one dimension fewer directions/positionsConsider only directions or positions on surfaces5View InterpolationDerived FrameReference FrameReference Frame• Create novel images by resampling photographso Reference images sample 5D plenoptic functionView Interpolation• Method:o Warp nearby reference images to novel viewpointo Blend warped imagesThis is just a morph where the warp is defined by pixel correspondences!R2R1NView Interpolation• How define warp for one view to another?o Use depth at pixel to project into scene, or ...R2NR1View Interpolation• How define warp for one view to another?o Use depth at pixel to project into scene, or ...View Interpolation• How define warp for one view to another?o Use depth at pixel to project into scene, or ...o Use pixel correspondencesView Interpolation• How define warp for one view to another?o Use depth at pixel to project into scene, or ...o Use pixel correspondences6View Interpolation• Finding pixel correspondenceso Coarse 3D modelo Sparse image featureso Depth at every pixel(Szeliski)LeftRightDisparityView Interpolation• Problems with view interpolation:o Changes in visibilityo Disocclusions(McMillan)Disocclusions• Partial solutions:o Fill holes by interpolating nearby pixels(McMillan)Disocclusions• Partial solutions:o Fill holes by interpolating nearby pixelso Use more photographs(Gortler)LumigraphImage-Based RepresentationsF(x, y, z, φ, θ, λ, t)7D6D5D4D3D2DIdealConsider only 3 frequencies (RGB)Consider rays through scene at one time instantConsider only lines through sceneConsider one dimension fewer directions/positionsConsider only directions or positions on surfacesLight Field / Lumigraph• If observer stays in free space, plenoptic function reduces to 4Do Exterior of the convex hull of an objecto Interior of an environment(Levoy96)F(r, α, φ, θ)7Representing a Light Field• Two-plane parameterization (4D)(Levoy96)Representing a Light Field(Levoy96)Two Interpretations of a Light Field(Levoy96)Creating a Light Field(Levoy96)Capturing a Light Field(Stanford University)Capturing a Light Field(Bennett Wilburn, Michal Sm ulski, Mark Horowitz)8Capturing a Light Field(Bennett Wilburn, Michal Sm ulski, Mark Horowitz)Capturing a Light Field(Bennett Wilburn, Michal Sm ulski, Mark Horowitz)Rendering a Light Field• Resampling problemo Interpolationo Avoid aliasing(Gortler96)Rendering a Light FieldVideo(Levo y & H anrahan)Rendering a Light FieldDemo(Levo y & H anrahan)Other IBR RepresentationsF(x, y, z, φ, θ, λ, t)7D6D5D4D3D2DIdealConsider only 3 frequencies (RGB)Consider rays through scene at one time instantConsider only lines through sceneConsider one dimension fewer directions/positionsConsider only directions or positions on surfaces9Sea of Images• Dense sampling with hemispherical camera moving in environment on eye-height planeRoboticCaptureDeviceCaptured viewpointsWalkthrough viewpoints(with Daniel (with Daniel AliagaAliaga))IBR Trade-offs• Advantageso Photorealistic - by definitiono Do not have to create 3D detailed modelo Do not have to do lighting simulationo Performance independent of scene• Disadvantageso Real-world scenes onlyo Difficult for dynamic sceneso Difficult to change rendering parameterso Difficult for scenes with specularities, etc.o Limited range of viewpointso Limited resolutionIBR Applications• Historical Site Preservation• Remote Education• Virtual TourismFrank Lloyd Wright Fallingwater House, PAThomas Jefferson Monticello, VAInside Independence Hall, Philadelphia, PAFutureComputerVisionComputerGraphicsImage-Based RenderingAnalyze &ReprojectSimulate(McMillan)Layered Depth Images• Multiple samples per pixel at different


View Full Document

Princeton COS 426 - Traditional Computer Graphics

Documents in this Course
Lecture

Lecture

35 pages

Lecture

Lecture

80 pages

Boids

Boids

25 pages

Exam 1

Exam 1

9 pages

Curves

Curves

4 pages

Lecture

Lecture

83 pages

Load more
Download Traditional Computer Graphics
Our administrator received your request to download this document. We will send you the file to your email shortly.
Loading Unlocking...
Login

Join to view Traditional Computer Graphics and access 3M+ class-specific study document.

or
We will never post anything without your permission.
Don't have an account?
Sign Up

Join to view Traditional Computer Graphics 2 2 and access 3M+ class-specific study document.

or

By creating an account you agree to our Privacy Policy and Terms Of Use

Already a member?