1 / 37

Binary Shading using Geometry and Appearance

This paper explores the integration of geometry and appearance in binary shading, presenting a variational rendering model for depicting scenes using limited color palettes. The authors discuss methods for enhancing geometric features through graph-based energy approaches, including results from various shading components like diffuse and specular effects. The framework accommodates both global and local feature control, making it suitable for interactive applications. Experiments demonstrate its effectiveness in generating images that exhibit a range of styles while maintaining a coherent visual representation.

alaire
Télécharger la présentation

Binary Shading using Geometry and Appearance

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. BinaryShadingusingGeometry and Appearance Bert BuchholzTamyBoubekeur Doug DeCarlo Marc Alexa Telecom ParisTech – CNRS RutgersUniversity TU Berlin Computer Graphics Forum Vol. 29, N. 6, 2010 PresentedatEurographics 2011

  2. BinaryDrawing Depictingscenesusing 2 colors

  3. BinaryRendering Geometry Camera 3D Scene Conversion Reflectance Lighting

  4. RelatedWork Half-Toning [Floyd and Steinberg 1976] [Ostromoukhov and Hersch 1995] • Image Binarization • [Mould and Kaplan 2008] • Local shadingoperators • [Vergne 2008] • Line Drawing • [DeCarlo 2003][Judd 2007]

  5. Analysis Per-pixel decision Contradictory criteria Geometry enhancement Low shading depiction power • Variational Rendering Model

  6. Variationalbinaryrendering

  7. VariationalBinaryShading • Deferred shading framework • Rendering data structured as a ST image graph • Edges energies derived from geometry and appearance • Final rendering as a min cut in the image graph

  8. DeferredShading Generate a set of arrays (render buffers) • GeometricProperties • AppearanceProperties Rasterization Ray Tracing

  9. Graph construction Standard Source-to-Sink Image Graph {V,E} Image Pixels Source (white) Sink (black) Image Connectivity Edges to Source Edges to SInk

  10. Appearance Contribution to the Graph • Through terminal weights on • AppearanceinitiateB/W segmentation • Experimentsusing : • Diffuse component • Specular component • Headlight component • Ambient Occlusion/Accessibilitycomponent Terminal Weights

  11. Appearance Graph Terms Global to local feature control using spatial averages Support size Local term: Sign:locallylighter or darker Weightedcombination to the global measure.

  12. Global versus Local Thresholding Global Local

  13. Geometry Contribution to the Graph • Modelledwithneighboredgeweights • Tailorsanisotropic, non-local B/W diffusion • Redistributes B/W values to enhancegeometricfeatures • Based on: • Viewdepth values • Viewdependentcurvature, estimated as screenspace normal derivatives [Judd 2007] Unit surface normal gatheredat Geodesic distance on the Gauss sphere

  14. GeometricTerm Local curvaturenormalization: Support size Yieldsneighbor edgeweights:

  15. Geometry Contribution Modulation

  16. Geometry Contribution Modulation Small components are successively connected

  17. Graph Minimum Cut • Max Flow Min Cut Theorem • Shortest split path in the graph →Feature size control through graph energy • Separate the graph in two components • Boykov and Kolmogorov implementation [2004]

  18. Cut performance • Rendering buffers canbe speed up usingrasterization • Measured on a Core2Duo 1.83GHz (single thread)

  19. Results

  20. BinaryShading • Global to local features • Appearance & geometry depiction • Large variety of style • Interactive control

  21. Comparison to Line Drawing BinaryShading Line Drawing

  22. Combinationwith Line Drawing

  23. Combinationwith Line Drawing RGBN Picture Rendering [Toler-Franklin 2007]

  24. Combinationwith Line Drawing

  25. Comparison to Thresholding Binary Rendering Diffuse Component Thresholding Thresholding Over Gaussian Filtering

  26. Comparison to Thresholding

  27. Comparison to Image Binarization (b,e) equivalent to [Mould and Kaplan 2008]

  28. On-goingwork Interactive control

  29. Interactive Control • Solution 1: slider-based, for experts • Accurate control • « Toomuch » control for novice users • Solution 2: painting interface

  30. Paint Interface • Supervised sparse B/W contraints • Derive the graph energy structure

  31. Paint Interface

  32. Limitations and future work

  33. Animation Works most of the time but… … most of straightforward solutions have some temporal coherencyfailure cases.

  34. K-colorRendering • Extend to multi-label cuts • Alternative energy minimizer • Lloyd relaxation/k-means • Mean Shift • k-Component cut • Vector Rendering

  35. conclusion

  36. BinaryShading as a Single Cut • Global solution • Local to global control • Appearance vs geometry control • High level control • For automaticbinarydrawingfrom 3D Scenes, decals, cut-out, etc…

  37. Thankyou BinaryShadingUsingGeometry and Appearance. Bert Buchholz, TamyBoubekeur, Doug DeCarlo and Marc Alexa Computer Graphics Forum Vol. 29, Nb. 6, 2010 PresentedatEurographics 2011

More Related