some recent graphics researchto inspire final projectsbryan klingner (klingner@cs)Procedural Modeling of Buildings - SIGGRAPH ‘06P. Mueller, P. Wonka, S. Haegler, A. Ulmer and L. Van Gool. 2006. Procedural Modeling of Buildings. In Proceedings of ACM SIGGRAPH 2006 / ACM Transactions on Graphics (TOG), ACM Press, Vol. 25, No. 3, pages 614-623.Procedural Modeling of Buildings - SIGGRAPH ‘06Production process:ß Rule-driven modification & replacement of shapesß Iteratively evolve a design by creating more andmore detailsß Sequential application (like Chomsky grammars)D. Hoiem, A.A. Efros, and M. Hebert, "Automatic Photo Pop-up", ACM SIGGRAPH 2005.Automatic Photo Pop-up - SIGGRAPH ‘05Automatic Photo Pop-up - SIGGRAPH ‘05Automatic Photo Pop-up - SIGGRAPH ‘05Salience-Preserving Color Removal - SIGGRAPH ‘05color imagegrayscalenew algorithmAmy A. Gooch, Sven C. Olsen, Jack Tumblin, Bruce Gooch. Color2Gray: Salience-Preserving Color Removal. ACM SIGGRAPH 2005.Salience-Preserving Color Removal - SIGGRAPH ‘05color grayscaleProblem: Isoluminant ColorsSalience-Preserving Color Removal - SIGGRAPH ‘05Salience-Preserving Color Removal - SIGGRAPH ‘05Salience-Preserving Color Removal - SIGGRAPH ‘05Original PhotoshopGrey Color2Greycolor Photoshop Gray Color2GraySalience-Preserving Color Removal - SIGGRAPH ‘05Physically-Based Animation and Modeling• Most things in graphics are animated by humans• Some things--like smoke, fire, and liquid--are too complex to feasibly animate realistically by hand• Instead, we use physical models of fluid flow, fracture, etc, cut corners, and render the result.A Method for Animating Viscoelastic Fluids - SIGGRAPH ‘05Animating Gases with Hybrid Meshes - SIGGRAPH ‘05Fluids in Deforming Meshes - SCA ‘05Fluid Animation with Dynamic Meshes - SIGGRAPH ‘06Simultaneous Coupling of Fluids and Deformable Bodies - SCA ‘06"Fast Separation of Direct and Global Components of a Scene using High Frequency Illumination,"S.K. Nayar, G. Krishnan, M. D. Grossberg, R. Raskar,ACM Trans. on Graphics (also Proc. of ACM SIGGRAPH),Jul, 2006.Fast Separation of Direct and Global Components of a Scene Using High Frequency Illumination - SIGGRAPH ‘06referred to as the direct component, Ld. The exact value of the directcomponent is determined by the BRDF of the surface point, whichcan be arbitrary1. For our separation method t o work, we assumethat each camera pixel can observe at most one significant scatter-ing event, i .e. two different source elements cannot produce a directcomponent along a camera pixel’s l ine of sight2.The remaining radiance measured by the camera pixel is referred toas the global component, Lg. In computer graphics, this term is typ-ically used to denote interreflections – l ight received by a surfacepoint after reflection by other scene points. Here, we are using amore general definition. In addition to interreflections, the global il-lumination received by the surface point may be due to volumetricscattering, subsurface scattering or even light diffusion by translu-cent surfaces (see Figure 2). The case of diffusion by a translucentsurface works similarly to interreflections. In the case of volumetricscattering, the global component arises from the illumination of thesurface point by light scattered f r om particles suspended i n a par-ticipating medium. In the case of subsurface scattering, the surfacepoint receives light from other points within the surface medium. Fi-nally, the global component also includes volumetric and subsurfaceeffects that may occur within the camera pixel’s field of view but out-side t he volume of intersection between the fields of view of the pixeland the source element that produces a significant scattering event atthe pixel. These are considered to be global effects as they are notsignificant scattering events caused by individual source elements3.In all cases, the total radiance measured at a camera pixel is the sumof the direct and global components:L = Ld+ Lg. (1)3.2 The Nature of the Light SourceIn our work, we restrict ourselves to the use of a single camera and asingle source. While we will use a point source to describe our sep-aration method, this is not a stri ct requirement. We only require thateach point in the scene be directly illuminated by at most one sourceelement. In other words, the light rays corresponding to the sourceelements should not intersect within the working volume of the setupused to perform the separation. Any source (point or extended) thatsatisfies this condition may be used.3.3 Separation using High Frequency IlluminationLet us assume that the scene in Figure 3(a) includes a single opaquesurface of arbitrary BRDF immersed in a non-scattering medium sothat the global component arises solely from interreflections. As wewill see, our analysis of this case is applicable to other phenomenasuch as subsurface and volumetric scattering.Let us divide the surface into a total of N patches, M of which aredirectly visible to the source. Each of these M visible patches corre-sponds to a single pixel of the source. We denote the radiance of thepatch i measured by the camera c as L[c, i], and its two componentsas Ld[c, i] and Lg[c, i], so that L[ c, i] = Ld[c, i] + Lg[c, i]. The globalcomponent of i due to interreflections from all other surface patchescan be written as:1Since, in practice, cameras and sources have fi nite resolutions, the directcomponent is the aggregate of all the scattering that occurs within the volumeof intersection between the fi elds of view of the camera pixel that observesthe surface point and the source element that illuminates it.2A scenario that violates this assumption is the case of a transparent (andyet reflective) surface in front of another surface, where a camera pixel’s lineof sight could produce two signifi cant scatterings due to different source ele-ments (pixels in the case of a projector).3This claim does not hold true when the source and the camera are co-located. In this special case, the fi eld of view of each camera pixel is illu-minated by a single source element and hence the volumetric and subsurfacescattering effects within the pixel’s fi eld of view will indeed be signifi cantscattering events and hence appear in the direct component.sourcecamerasurfaceisourcecamerasurfacei(a)sourcecamerasurfaceisourcecamerasurfacei(b)Figure 3: (a) A simple scenario where the radiance of each patchi includes a direct component due to
View Full Document
Unlocking...