350 likes | 793 Views
Paint By Numbers : Abstract Image Representation Paul Haeberli Silicon Graphics Computer Systems ACM SIGGRAPH Computer graphics Proceeding of 17 th , 1990 Outline Motivation Overview Painting Techniques Stroke Attributes Operations on Paintings Advanced Techniques Spice for Images
E N D
Paint By Numbers : Abstract Image Representation Paul Haeberli Silicon Graphics Computer Systems ACM SIGGRAPH Computer graphics Proceeding of 17th, 1990
Outline • Motivation • Overview • Painting Techniques • Stroke Attributes • Operations on Paintings • Advanced Techniques • Spice for Images • Conclusion • Further work
Motivation • Producing images indistinguishable from photograph • Graphic designer’s choice • Photorealistic, Not always the best choice • “How much use is a photograph to mechanics when they already have the real thing on front of them?”[Lansdown and Schofield] • Visual effect intended by designer [Lansdown and Schofield] J.Lansdown and S.Schofield. Expressive rendering: A review of nonphotorealistic techniques. IEEE ComputerGraphics and Applications, 1995.
Overview • Alternative to photorealism • Painterly rendering • Creation of artistic, stylized and abstract images • Impressionistic painting • Brush stroke control • User interactive system
Operation • Information from source image • Cursor across the canvas • Sampling color from image • Paint a brush stroke • Location • Color • Size • Direction • Shape
Stroke Location • Stochastic distribution around cursor • Example of interactive particle system[Reeves] [Reeves] William T. Reeves and Ricki Blau, “Approximate and probabilistic algorithm for shading and rendering structured particle systems”, Computer Graphics, 1985.
Stroke Color • RGB and alpha value • Time limitation to pick new color • “put-that-color-there” procedure[Lewis] • Restrict to small number of color [Lewis] John-Peter Lewis, “Texture Synthesis for Digital Painting”, Computer Graphics, 1984.
Stroke Size and Orientation • Size • Control by cursor speed • Easy to create rough representation • Control by arrow keys • Orientation • Direction of cursor • Mouse gesture • Image gradient
Stroke Shape • Shape • Significant influenceto final painting • Circle, rectangle, line,scattering of points,polygon, cone, user-defined shape
Painting Example Diagonal stroke Pointillist representation
Painting Example • • Cone shape brush • • Voronoi diagram polygon resterizing hardware• rendering of cones to construct 2D Voronoi diagrams of points Cone shape
Painting Description • Painting as an ordered set of strokes • Containing stroke information • Operations on paintings • Transform painting into RGB images • Unary operation – scaling, sorting, adding noise, etc. • Binary operation – interpolation, extrapolation, animation, etc
Brush Direction • Brush direction using second image
Edge drawing • Edge drawing using luminance gradient
Texture mapping • Brush texture mapping
Sampling Geometry • Sampling geometry using Ray-tracing
Approximation • Approximation using Relaxation
Edge enhancement • “Pushing edge” • More explicit depth relationship • Using unsharp masking
Color enhancement • Increase saturation • Lum = 0.3*R + 0.59*G + 0.11*B • Extrapolation
Color restriction and Background Color • Color restriction • Limited color for overall harmony and unity • Color quantization of source image • Noise for no contouring • Background cover • Unity and integrity • Color perception
Conclusion • A system for producing abstract image • The first experiment on NPR • Making stylized abstract image • Interactive processing • Motivations for future works • Media • Method
Inspiration • Haeberli’s work • The first work on NPR • Some issues • Stroke methods, attributes, animation, etc • After Haeberli’s work • Considerable works on painting system • Two large categorization • Digital analogues • Automatic stroke
Digital analogues • Salisbury et al. 1994 • Pen and ink • Curtis et al. 1997 • Watercolor • Sousa et al. 1999 • Pencil drawing
Automatic stroke • Litwinowicz, 1997 • Hertzmann, 1998 • Shiraishi and Yamaguchi, 2000 • Only global effect by user • Brush size or shape
Stroke Dimension • Dana, 1996 • 8D dimension for social visualization[Dana] • To show social data Animation • Meier, 1996 • Using particle rendering methods