SlideShare a Scribd company logo
Relief: A Modeling by Drawing Tool David Bourguignon 1   Raphaëlle Chaine 2 Marie-Paule Cani 3   George Drettakis 4 1 Princeton University / INRIA Rocquencourt  2 LIRIS / CNRS / UCBL 3 GRAVIR / INP Grenoble  4 REVES / INRIA Sophia-Antipolis
Outline Motivation Previous Work Tool Workflow Reconstruction Adaptive Sampling & Depth Inference Tool Interface Results
On Users Most people draw Writing alternative Few people sculpt Play-Doh days long gone Materials difficult to handle
Goals Use 2D tools to perform 3D operations
Goals Use 2D tools to perform 3D operations Model global and local surface
Goals Use 2D tools to perform 3D operations Model global and local surface Input: just plain strokes
Goals Use 2D tools to perform 3D operations Model global and local surface Input: just plain strokes Output: triangle mesh
Outline Motivations Previous Work Tool Workflow Reconstruction Adaptive Sampling & Depth Inference Tool Interface Results
Previous Work Depth painting [Williams, 1990] +
Previous Work Gradient editing [van Overveld, 1996]
Previous Work Maya 6.0 Artisan [Alias, 2004]
Outline Motivations Previous Work Tool Workflow Reconstruction Adaptive Sampling & Depth Inference Tool Interface Results
Tool Workflow First step: drawing input Displacement map mid-grey = 0 white > 0 black < 0 Model of 3D sphere Pencil  Brush
Tool Workflow First step: drawing Displacement map 2D shape boundary (in green) defines drawing mask
Tool Workflow First step: drawing Displacement map 2D shape boundary Displacement regions (from 2 maps)
Tool Workflow Second step: modeling Displace existing vertices
Tool Workflow Second step: modeling Displace existing vertices Create new surface  patch
Tool Workflow Changing viewpoint Modeling by drawing Changing viewpoint
Reconstruction Based on evolving pseudo-manifold [Chaine, 2003]
Reconstruction Based on evolving pseudo-manifold [Chaine, 2003] Satisfy our requirements Arbitrary number of connected components
Reconstruction Based on evolving pseudo-manifold [Chaine, 2003] Satisfy our requirements Arbitrary number of connected components Handle points off shape boundary
Reconstruction Based on evolving pseudo-manifold [Chaine, 2003] Satisfy our requirements Arbitrary number of connected components Handle points off shape boundary Interactive (5k points per second)
2D reconstruction Start: pseudo-curve lies on oriented edges of Delaunay triangulation
2D reconstruction During: pseudo-curve evolves as long as oriented Gabriel criterion is not met
2D reconstruction Stop: topologically consistent set of oriented edges
Sampling and Depth Adaptive sampling Displacement map Pencil and brush data in color buffer Color buffer
Sampling and Depth Adaptive sampling Displacement map Approximate disp. map sampled at existing  vertices
Sampling and Depth Adaptive sampling Displacement map (D) Vertex-Sampled disp.  map (V) Error map  E = 1 – ABS(D – V) Arbitrary error value
Sampling and Depth Adaptive sampling Displacement map Approximate disp. map Error map Sampling [Alliez, 2002]
Sampling and Depth Adaptive sampling Depth inference Identify surface vertices Vertices ID buffer
Sampling and Depth Adaptive sampling Depth inference Identify surface vertices Assign depth values Depth buffer
Sampling and Depth Adaptive sampling Depth inference Identify surface vertices Assign depth values Infer depth values from existing surface by depth propagation
Outline Motivations Previous Work Tool Workflow Reconstruction Adaptive Sampling & Depth Inference Tool Interface Results
Tool Interface Hole marks Comic books production Hole marks Stone #3 (Avalon Studios)
Tool Interface Hole marks Comic books production Our system Hole mark
Tool Interface Video: Basic interface
Tool Interface Blobbing Drawing White shading Distance field Height field Surface
Tool Interface Depth modes (chosen by menu) Modeling “at depth” Depth inference Frisket mode
Video Modeling a tree Paper sketch 3D model obtained with Relief
Outline Motivations Previous Work Tool Workflow Reconstruction Adaptive Sampling & Depth Inference Tool Interface Results
Results Models (1k to 4k points)
Discussion Intuitive shading convention
Discussion Intuitive shading convention Problems with drawing metaphor No continuous visual feedback Provide two modes
Discussion Intuitive shading convention Problems with drawing metaphor No continuous visual feedback Difficult to obtain continuous shading Provide higher-level drawing tools
Conclusion Modeling by drawing, but imprecise
Conclusion Modeling by drawing, but imprecise Future work Speedup with local 3D reconstruction
Conclusion Modeling by drawing, but imprecise Future work Speedup with local 3D reconstruction Improve depth inference
Conclusion Modeling by drawing, but imprecise Future work Speedup with local 3D reconstruction Improve depth inference Image-space and object-space sampling
Acknowledgements This work has been performed while the first author was a visiting research fellow at Princeton University, supported by an INRIA post-doctoral fellowship. Many people have indirectly contributed to it. We would like to thank: Adam Finkelstein, Szymon Rusinkiewicz, Jason Lawrence, Pierre Alliez, Mariette Yvinec, Laurence Boissieux, Laure Heïgéas, Laks Raghupathi, Olivier Cuisenaire, Bingfeng Zhou.
 

More Related Content

PPT
Drawing For Illustration And Annotation In 3D
PPTX
GRPHICS02 - Creating 3D Graphics
PPTX
Understanding neural radiance fields
PPTX
study Diffusion Curves: A Vector Representation for Smooth-Shaded Images
PPTX
Datt2500 week3
DOC
Photoshop Tutorials Trafic Light
PPT
UIUC CS 498 - Computational Photography - Final project presentation
PPTX
Image Stitching for Panorama View
Drawing For Illustration And Annotation In 3D
GRPHICS02 - Creating 3D Graphics
Understanding neural radiance fields
study Diffusion Curves: A Vector Representation for Smooth-Shaded Images
Datt2500 week3
Photoshop Tutorials Trafic Light
UIUC CS 498 - Computational Photography - Final project presentation
Image Stitching for Panorama View

What's hot (18)

PPTX
Datt2500 week5
PDF
Image Printing Based on Halftoning
PPT
Raster vs Vector
PDF
A complete interior design solution with diminished reality
PPTX
Image enhancement
PPTX
Datt 2500 week 10
PPTX
Raster vs Vector files
PPTX
Mathematical operations in image processing
PDF
Raster vs vector
PDF
JonathanWestlake_ComputerVision_Project1
PPTX
2012 13 raster vs vector
PPTX
3D theory: geometry, extrusion, co-ordinate theory
PPTX
Digital Image processing
PPTX
DSP presentation_latest
DOCX
Raster and vector
PPT
Raster Vs. Vector Presentation1
PPTX
Dilation and erosion
PDF
Improved Alpha-Tested Magnification for Vector Textures and Special Effects
Datt2500 week5
Image Printing Based on Halftoning
Raster vs Vector
A complete interior design solution with diminished reality
Image enhancement
Datt 2500 week 10
Raster vs Vector files
Mathematical operations in image processing
Raster vs vector
JonathanWestlake_ComputerVision_Project1
2012 13 raster vs vector
3D theory: geometry, extrusion, co-ordinate theory
Digital Image processing
DSP presentation_latest
Raster and vector
Raster Vs. Vector Presentation1
Dilation and erosion
Improved Alpha-Tested Magnification for Vector Textures and Special Effects
Ad

Similar to Relief: A Modeling By Drawing Tool (20)

PPTX
Collaborative 3D Modeling by the Crowd
PDF
M sc thesis proposal v4
PPT
Interactive Stereoscopic Rendering for Non-Planar Projections (GRAPP 2009)
PPTX
3D Image visualization
PPTX
Stylized shading - FMX 2012
PDF
lecture_16_jiajun.pdf
PPT
Build Your Own 3D Scanner: 3D Scanning with Swept-Planes
PDF
Faire de la reconnaissance d'images avec le Deep Learning - Cristina & Pierre...
PPTX
Detection ofs Signlanguageminorppt1.pptx
PPTX
A Dynamic Noise Primitive for Coherent Stylization, EGSR 2010
PPT
Non-Planar Projections (GRAPP 2008)
PDF
How to use transfer learning to bootstrap image classification and question a...
PPT
Real-Time Volumetric Tests (EG 2008)
PDF
Android Apps Training - Day Four (Design)
PPTX
OReilly AI Transfer Learning
PDF
2013 Lecture 5: AR Tools and Interaction
PPTX
AutoCAD drawing layout using layout tool
PPT
Advanced Lighting for Interactive Applications
Collaborative 3D Modeling by the Crowd
M sc thesis proposal v4
Interactive Stereoscopic Rendering for Non-Planar Projections (GRAPP 2009)
3D Image visualization
Stylized shading - FMX 2012
lecture_16_jiajun.pdf
Build Your Own 3D Scanner: 3D Scanning with Swept-Planes
Faire de la reconnaissance d'images avec le Deep Learning - Cristina & Pierre...
Detection ofs Signlanguageminorppt1.pptx
A Dynamic Noise Primitive for Coherent Stylization, EGSR 2010
Non-Planar Projections (GRAPP 2008)
How to use transfer learning to bootstrap image classification and question a...
Real-Time Volumetric Tests (EG 2008)
Android Apps Training - Day Four (Design)
OReilly AI Transfer Learning
2013 Lecture 5: AR Tools and Interaction
AutoCAD drawing layout using layout tool
Advanced Lighting for Interactive Applications
Ad

Recently uploaded (20)

PDF
Diabetes mellitus diagnosis method based random forest with bat algorithm
PDF
NewMind AI Weekly Chronicles - August'25-Week II
PPTX
Programs and apps: productivity, graphics, security and other tools
PDF
Accuracy of neural networks in brain wave diagnosis of schizophrenia
PPTX
KOM of Painting work and Equipment Insulation REV00 update 25-dec.pptx
PDF
Unlocking AI with Model Context Protocol (MCP)
PDF
Assigned Numbers - 2025 - Bluetooth® Document
PPTX
Tartificialntelligence_presentation.pptx
PDF
Univ-Connecticut-ChatGPT-Presentaion.pdf
PPTX
Spectroscopy.pptx food analysis technology
PDF
Mushroom cultivation and it's methods.pdf
PPTX
1. Introduction to Computer Programming.pptx
PPTX
Machine Learning_overview_presentation.pptx
PDF
Encapsulation theory and applications.pdf
PDF
7 ChatGPT Prompts to Help You Define Your Ideal Customer Profile.pdf
PDF
Building Integrated photovoltaic BIPV_UPV.pdf
PDF
Spectral efficient network and resource selection model in 5G networks
PDF
Blue Purple Modern Animated Computer Science Presentation.pdf.pdf
PPTX
TLE Review Electricity (Electricity).pptx
PDF
Per capita expenditure prediction using model stacking based on satellite ima...
Diabetes mellitus diagnosis method based random forest with bat algorithm
NewMind AI Weekly Chronicles - August'25-Week II
Programs and apps: productivity, graphics, security and other tools
Accuracy of neural networks in brain wave diagnosis of schizophrenia
KOM of Painting work and Equipment Insulation REV00 update 25-dec.pptx
Unlocking AI with Model Context Protocol (MCP)
Assigned Numbers - 2025 - Bluetooth® Document
Tartificialntelligence_presentation.pptx
Univ-Connecticut-ChatGPT-Presentaion.pdf
Spectroscopy.pptx food analysis technology
Mushroom cultivation and it's methods.pdf
1. Introduction to Computer Programming.pptx
Machine Learning_overview_presentation.pptx
Encapsulation theory and applications.pdf
7 ChatGPT Prompts to Help You Define Your Ideal Customer Profile.pdf
Building Integrated photovoltaic BIPV_UPV.pdf
Spectral efficient network and resource selection model in 5G networks
Blue Purple Modern Animated Computer Science Presentation.pdf.pdf
TLE Review Electricity (Electricity).pptx
Per capita expenditure prediction using model stacking based on satellite ima...

Relief: A Modeling By Drawing Tool

  • 1. Relief: A Modeling by Drawing Tool David Bourguignon 1 Raphaëlle Chaine 2 Marie-Paule Cani 3 George Drettakis 4 1 Princeton University / INRIA Rocquencourt 2 LIRIS / CNRS / UCBL 3 GRAVIR / INP Grenoble 4 REVES / INRIA Sophia-Antipolis
  • 2. Outline Motivation Previous Work Tool Workflow Reconstruction Adaptive Sampling & Depth Inference Tool Interface Results
  • 3. On Users Most people draw Writing alternative Few people sculpt Play-Doh days long gone Materials difficult to handle
  • 4. Goals Use 2D tools to perform 3D operations
  • 5. Goals Use 2D tools to perform 3D operations Model global and local surface
  • 6. Goals Use 2D tools to perform 3D operations Model global and local surface Input: just plain strokes
  • 7. Goals Use 2D tools to perform 3D operations Model global and local surface Input: just plain strokes Output: triangle mesh
  • 8. Outline Motivations Previous Work Tool Workflow Reconstruction Adaptive Sampling & Depth Inference Tool Interface Results
  • 9. Previous Work Depth painting [Williams, 1990] +
  • 10. Previous Work Gradient editing [van Overveld, 1996]
  • 11. Previous Work Maya 6.0 Artisan [Alias, 2004]
  • 12. Outline Motivations Previous Work Tool Workflow Reconstruction Adaptive Sampling & Depth Inference Tool Interface Results
  • 13. Tool Workflow First step: drawing input Displacement map mid-grey = 0 white > 0 black < 0 Model of 3D sphere Pencil Brush
  • 14. Tool Workflow First step: drawing Displacement map 2D shape boundary (in green) defines drawing mask
  • 15. Tool Workflow First step: drawing Displacement map 2D shape boundary Displacement regions (from 2 maps)
  • 16. Tool Workflow Second step: modeling Displace existing vertices
  • 17. Tool Workflow Second step: modeling Displace existing vertices Create new surface patch
  • 18. Tool Workflow Changing viewpoint Modeling by drawing Changing viewpoint
  • 19. Reconstruction Based on evolving pseudo-manifold [Chaine, 2003]
  • 20. Reconstruction Based on evolving pseudo-manifold [Chaine, 2003] Satisfy our requirements Arbitrary number of connected components
  • 21. Reconstruction Based on evolving pseudo-manifold [Chaine, 2003] Satisfy our requirements Arbitrary number of connected components Handle points off shape boundary
  • 22. Reconstruction Based on evolving pseudo-manifold [Chaine, 2003] Satisfy our requirements Arbitrary number of connected components Handle points off shape boundary Interactive (5k points per second)
  • 23. 2D reconstruction Start: pseudo-curve lies on oriented edges of Delaunay triangulation
  • 24. 2D reconstruction During: pseudo-curve evolves as long as oriented Gabriel criterion is not met
  • 25. 2D reconstruction Stop: topologically consistent set of oriented edges
  • 26. Sampling and Depth Adaptive sampling Displacement map Pencil and brush data in color buffer Color buffer
  • 27. Sampling and Depth Adaptive sampling Displacement map Approximate disp. map sampled at existing vertices
  • 28. Sampling and Depth Adaptive sampling Displacement map (D) Vertex-Sampled disp. map (V) Error map E = 1 – ABS(D – V) Arbitrary error value
  • 29. Sampling and Depth Adaptive sampling Displacement map Approximate disp. map Error map Sampling [Alliez, 2002]
  • 30. Sampling and Depth Adaptive sampling Depth inference Identify surface vertices Vertices ID buffer
  • 31. Sampling and Depth Adaptive sampling Depth inference Identify surface vertices Assign depth values Depth buffer
  • 32. Sampling and Depth Adaptive sampling Depth inference Identify surface vertices Assign depth values Infer depth values from existing surface by depth propagation
  • 33. Outline Motivations Previous Work Tool Workflow Reconstruction Adaptive Sampling & Depth Inference Tool Interface Results
  • 34. Tool Interface Hole marks Comic books production Hole marks Stone #3 (Avalon Studios)
  • 35. Tool Interface Hole marks Comic books production Our system Hole mark
  • 36. Tool Interface Video: Basic interface
  • 37. Tool Interface Blobbing Drawing White shading Distance field Height field Surface
  • 38. Tool Interface Depth modes (chosen by menu) Modeling “at depth” Depth inference Frisket mode
  • 39. Video Modeling a tree Paper sketch 3D model obtained with Relief
  • 40. Outline Motivations Previous Work Tool Workflow Reconstruction Adaptive Sampling & Depth Inference Tool Interface Results
  • 41. Results Models (1k to 4k points)
  • 43. Discussion Intuitive shading convention Problems with drawing metaphor No continuous visual feedback Provide two modes
  • 44. Discussion Intuitive shading convention Problems with drawing metaphor No continuous visual feedback Difficult to obtain continuous shading Provide higher-level drawing tools
  • 45. Conclusion Modeling by drawing, but imprecise
  • 46. Conclusion Modeling by drawing, but imprecise Future work Speedup with local 3D reconstruction
  • 47. Conclusion Modeling by drawing, but imprecise Future work Speedup with local 3D reconstruction Improve depth inference
  • 48. Conclusion Modeling by drawing, but imprecise Future work Speedup with local 3D reconstruction Improve depth inference Image-space and object-space sampling
  • 49. Acknowledgements This work has been performed while the first author was a visiting research fellow at Princeton University, supported by an INRIA post-doctoral fellowship. Many people have indirectly contributed to it. We would like to thank: Adam Finkelstein, Szymon Rusinkiewicz, Jason Lawrence, Pierre Alliez, Mariette Yvinec, Laurence Boissieux, Laure Heïgéas, Laks Raghupathi, Olivier Cuisenaire, Bingfeng Zhou.
  • 50.