Some thoughts based on our chosen goal

1. The structure or edges as our first-priority level: Small dots might be good. Density should be based on the budget.

(1) Old Halftoning scheme is not good for displaying different levels based on required number of stipples. Scaling the original image is a fast way to keep partial structure.

(2) Still based on halftoning, to find a good way to select important dots.

(3) Modify halftoning to have it matched with varying-size dots

(4) If using edge or gradient information to manage the orders, ....

2. Tone can be represented a little as an enhancement.

(1)After having good structure for stippling, tone as an enhancement

(2) A little would work.

3. Some global preservation: intensity compared with the original image and the structure

To have control over the possible number of dots

Fewer dots reduce intensity???? Does this prove our scaling method?

4. Some local properties:

(1) dot sizes:

Big dots are rare except typeIII.

Varying sizes are rare except typeIII.

a little change of size would be fine.

Thus, the range of dot sizes should be small(floating points: twice). It should based on the intensity.

Some problems from varying sizes (ramp, uniform, and edges): e.g. for a uniform region, if based on the black coverage of the dots to determine the intensity of an area, bigger dots need fewer number to cover an region than smaller dots need, which makes region with smaller dots having more blank space between dots. Will the visualization from smaller dots appear lighter even the value of intensity is same?

(2) intensity: besides black and white, introduce greyscale as an makeup. It is rare, but can be as an additional effect.

(3) density: based on the number of dots in a unit area? or based on the blank space.

Should be evenly spaced

(4) dot shape

5. Time tolerated

Stippling from artists

A site related to our goal: http://actingwhite.blogspot.com/search/label/Politics

Here I grabbed some art work from artist to classify different types of stippling. TypeI and Type IV are main streams. Our goal is similar to No. 2(TypeII) and No. 6 effects(TypeV). Right now some of results are similar to No. 7.

stippling based on scaling

Different scaling function

Small modifications for Stippling

Save the stipples as .svg, .eps, and fig files.

Scale each intensity to do the halftoning and represent them as stipples with radius 2
Three-level display

First try for displaying patterns(maybe useless maybe not)

#define DEF_MAXIMUM_COST_SPREAD 6000

#define DEF_MINIMUM_LENGTH_DENTRITE 0

#define DEF_WEIGHT_GRADIENT 20

Not well-defined for shapes and well-defined for sizes

It is possible to show patterns, but the quality of the patterns is questionable. And did not see many advantages over the texture-mapping way. But still interesting for me.

Stippling with sizes from average intensities after CAH

Do CAH and calculate the sizes of each stipple based on the average intensities. Redraw the stipples. If the new stipple covers the old pixels after CAH, the old pixels won't be stipples any more.

(A) It seems it keeps some edges after half procedure.

(A) Half pixels are processed.
(B)Zoom in from (C)

(C) All pixels drawn as stipples

Stippling with sizes from merged dots after CAH

Do the merging based on the CAH and redraw the stipples based on the sizes of the merged dots.

Problems:

1) Not follow the edges well. Might be improved by adjusting the center position when redrawing.

2) Tone is not good, either.

Good:

1) It seems it is promising if we can find a good selection from CAH

(A)

Do not draw the small stipples.

(B)

All of stipples

Three-level stippling similar effects to hightening

Kind of hightening effects: a very powerful expression for me.

It looks more interesting for me now to create three-level effects such as hightening.

Stippling based on grayness

After halftoning, there are nearly half number of black dots, nearly half number of white dots, and very few grey dots.

Halftoning to Stippling

Problems:

1. Not good at density control. From this view, the growing mask seems more promising.

======================

Scaling errors

Stippling
Halftoning
============================
(Half number of dots are used)

StipplingHalftoning
=========================================

Stippling
Halftoning
==============================================

Digital facial engraving

http://www.iro.umontreal.ca/~ostrom/publications/abstracts.html
Victor Ostromoukhov. Digital Facial Engraving, In Proceedings of SIGGRAPH'99, in Computer Graphics Proceedings, Annual Conference Series, pp. 417-424, 1999.

Gouraud shading lines

Test illuminating lines

The height field for a sphere is created by calculating the distance between the light and the vertices of the sphere. The lines are shown as contour lines based on different light position. Next step is to try to make the contour line continuous or long.

Random heights for a sphere.

a sphere
a sphere
a sphere
all illuminating lines
Triangulation for irregular PSLG

Three styles