Non-Photorealistic Rendering

Transcription

1 Computer Graphics I Lecture 22 Non-Photorealistic Rendering November 18, 2003 Doug James Carnegie Mellon University Pen-and-Ink Illustrations Painterly Rendering Cartoon Shading Technical Illustrations

2 Goals of Computer Graphics Traditional: Photorealism Sometimes, we want more Cartoons Artistic expression in paint, pen-and-ink Technical illustrations Scientific visualization 12/2/ Graphics I 2

3 Non-Photorealistic Rendering A means of creating imagery that does not aspire to realism - Stuart Green Cassidy Curtis 1998 David Gainey 12/2/ Graphics I 3

4 Some NPR Categories Pen-and-Ink illustration Techniques: cross-hatching, outlines, line art,etc. Painterly rendering Styles: impressionist, expressionist, pointilist, etc. Cartoons Effects: cartoon shading, distortion, etc. Technical illustrations Characteristics: Matte shading, edge lines, etc. Scientific visualization Methods: splatting, hedgehogs, etc. 12/2/ Graphics I 4

5 Emergence of NPR 2D Paint (Pixel Oriented) Bitmap paint systems 2D Paint (Brush Oriented) User intervention 2D/2.5D Paint Post-Processing Automatically generated from augmented images 3D Renderers Automatically generated based on 3D data 3D Photorealistic Renderers Traditional Computer Graphics 12/2/ Graphics I 5

6 Outline Pen-and-Ink Illustrations Painterly Rendering Cartoon Shading Technical Illustrations 12/2/ Graphics I 6

7 Pen-and-Ink Illustrations Strokes Curved lines of varying thickness and density Texture Character conveyed by collection of strokes Tone Perceived gray level across image or segment Outline Boundary lines that disambiguate structure 12/2/ Graphics I 7

8 Pen-and-Ink Examples Winkenbach and Salesin /2/ Graphics I 8

9 Rendering Polygonal Surfaces 3D Model Lighting Camera How much 3D information do we preserve? Visible Polygons Procedural Stroke Texture Stroke Clipping Outline Drawing 12/2/ Graphics I 9

10 Strokes and Stroke Textures Stroke generated by moving along straight path Stroke perturbed by Waviness function (straightness) Pressure function (thickness) Collected in stroke textures Tone dependent Resolution dependent Orientation dependent How automatic are stroke textures 12/2/ Graphics I 10

11 Stroke Texture Examples Winkenbach and Salesin /2/ Graphics I 11

12 Prioritized Stroke Textures Technique for limiting human intervention Collection of strokes with associated priority When rendering First draw highest priority only If too light, draw next highest priority, etc. Stop if proper tone is achieved Procedural stroke textures Support scaling Also applies to non-procedural stroke textures 12/2/ Graphics I 12

13 Stroke Texture Operations Scaling Changing Viewing Direction (Anisotropic) 12/2/ Graphics I 13

14 Indication Selective addition of detail Difficult to automate User places detail segments interactively 12/2/ Graphics I 14

15 Indication Example Bold strokes indicate detail segments With indication Without indication 12/2/ Graphics I 15

16 Outlines Boundary or interior outlines Accented outlines for shadowing and relief Dependence on viewing direction Suggest shadow direction 12/2/ Graphics I 16

17 Rendering Parametric Surfaces Stroke orientation and density Place strokes along isoparameter lines Choose density for desired tone tone = width / spacing v u 12/2/ Graphics I 17

18 Stroke Width Adjust stroke width retain uniform tone Winkenbach and Salesin /2/ Graphics I 18

19 Parametric Surface Example Constant-density hatching Smooth shading with single light Longer smoother strokes for glass Environment mapping Update reflection coefficient Standard rendering techniques are still important! 12/2/ Graphics I 19

20 Parametric Surface Example Winkenbach and Salesin /2/ Graphics I 20

21 Orientable Textures Inputs Grayscale image to specify desired tone Direction field Stroke character Output Stroke shaded image Salisbury et al /2/ Graphics I 21

22 Orientable Stroke Texture Example Salisbury et al /2/ Graphics I 22

23 Illustrating Smooth Surfaces [Hertzmann & Zorin, SIGGRAPH 2000] 12/2/ Graphics I 23

24 Animating Traditional Pencil Drawings From SIGGRAPH 2003 course notes on NPR; Daniel Teece, Walt Disney Feature Animation 12/2/ Graphics I 24

25 Animating Traditional Pencil Drawings From SIGGRAPH 2003 course notes on NPR; Daniel Teece, Walt Disney Feature Animation 12/2/ Graphics I 25

26 Outline Pen-and-Ink Illustrations Painterly Rendering Cartoon Shading Technical Illustrations 12/2/ Graphics I 26

27 Painterly Rendering Physical simulation User applies brushstrokes Computer simulates media Automatic painting User provides input image or 3D model User specifies painting parameters Computer generates all strokes Subject to controversy 12/2/ Graphics I 27

28 Physical Simulation Example Curtis et al. 1997, Computer Generated Watercolor 12/2/ Graphics I 28

29 Computer-Generated Watercolor Complex physical phenomena for artistic effect Build simple approximations Paper generation as random height field Simulated effects 12/2/ Graphics I 29

30 Fluid Simulation Use water velocity, viscosity, drag, pressure, pigment concentration, paper gradient Paper saturation and capacity Discretize and use cellular automata 12/2/ Graphics I 30

31 Interactive Painting Simulation in progress Finished painting 12/2/ Graphics I 31

32 Automatic Painting Example Hertzmann /2/ Graphics I 32

33 Automatic Painting from Images Start from color image: no 3D information Paint in resolution-based layers Blur to current resolution Select brush based on current resolution Find area of largest error compared to real image Place stroke Increase resolution and repeat Layers are painted coarse-to-fine Styles controled by parameters 12/2/ Graphics I 33

34 Layered Painting Blurring Adding detail with smaller strokes 12/2/ Graphics I 34

35 Brush Strokes Start at point of maximal error Calculate difference between original image and image painted so far Direction perpendicular to gradient Stroke tends to follow equally shaded area Stopping criteria Difference between brush color and original image color exceeds threshold Maximal stroke length reached 12/2/ Graphics I 35

36 Longer Brush Strokes For longer, curved brush strokes Repeat straight line algorithm Stop, again on length or difference threshold Use anti-aliased cubic B-spline 12/2/ Graphics I 36

37 Painting Styles Style determined by parameters Approximation threshold Brush sizes Curvature filter Blur factor Minimum and maximum stroke lengths Opacity Grid size Color jitter Encapsulate parameter settings as style 12/2/ Graphics I 37

38 Some Styles Impressionist No random color, 4 stroke length 16 Brush sizes 8, 4, 2; approximation threshold 100 Expressionist Random factor 0.5, 10 stroke length 16 Brush sizes 8, 4, 2; approximation threshold 50 Pointilist Random factor ~0.75, 0 stroke length 0 Brush sizes 4, 2; approximation threshold 100 Not convincing to artists 12/2/ Graphics I 38

39 Style Examples 12/2/ Graphics I 39

40 Outline Pen-and-Ink Illustrations Painterly Rendering Cartoon Shading Technical Illustrations 12/2/ Graphics I 40

41 Cartoon Shading Shading model in 2D cartoon Use material color and shadow color Present lighting cues, shape, and context Stylistic Used in many animated movies Developing real-time techniques for games 12/2/ Graphics I 41

42 Cartoon Shading as Texture Map Apply shading as 1D texture map Carl Marshall 2000 u=n L 12/2/ Graphics I 42

43 Shading Variations Flat shading Shadow Shadow + highlight 12/2/ Graphics I 43

44 Outline Pen-and-Ink Illustrations Painterly Rendering Cartoon Shading Technical Illustrations 12/2/ Graphics I 44

45 Technical Illustrations Level of abstraction Accent important 3D properties Dimish or eliminate extraneous details Do not represent reality Ruppel 1995 Photo 12/2/ Graphics I 45

46 Conventions in Technical Illustrations Black edge lines Cool to warm shading colors Single light source; shadows rarely used 12/2/ Graphics I 46

47 Technical Illustration Example Phong shading Metal shading (anisotropic) Edge lines Tone shading (cool to warm shift) 12/2/ Graphics I 47

48 Scientific Visualization Effective visualization of large, multidimensional datasets Turk & Banks, Image-Guided Streamline Placement, SIGGRAPH 96 12/2/ Graphics I 48

49 How to evaluate/define? Smart graphics design from user s perspective with data? HCI, AI, Perceptual studies Artistic graphics beyond imitating a way to create art work how to assess? 12/2/ Graphics I 49

50 Art-Based Rendering of Fur, Grass and Trees [Kowalski et al., SIGGRAPH 99] 12/2/ Graphics I 50

51 Art-Based Rendering of Fur, Grass and Trees [Kowalski et al., SIGGRAPH 99] 12/2/ Graphics I 51

52 Doug DeCarlo, Anthony Santella. Stylization and Abstraction of Photographs In SIGGRAPH /2/ Graphics I 52

53 12/2/ Graphics I 53

54 12/2/ Graphics I 54

55 Image Analogies [Hertzmann et al., SIGGRAPH 2001]? 12/2/ Graphics I 55

56 Image Analogies [Hertzmann et al., SIGGRAPH 2001] 12/2/ Graphics I 56

57 Image Analogies [Hertzmann et al., SIGGRAPH 2001] 12/2/ Graphics I 57

58 Carl Marshall 12/2/ Graphics I 58