Statistical Data Mining: A Short Course for the Army Conference on Applied Statistics

1
Statistical Data Mining A Short Course for the
Army Conference on Applied Statistics

• Edward J. Wegman
• George Mason University
• Jeffrey L. Solka
• Naval Surface Warfare Center

2
Color Theory Color Design Graphic Design
3
Color Theory
Human vision relies on light sensitive cells in
the retina of the eye. There are two basic kinds
of sensors. These are rods and cones. Rods are
cells which can work at very low intensity, but
cannot resolve sharp images or color. Cones are
cells that can resolve sharp images and color,
but require much higher light levels to work.
There are about 107 foveal cones. The combined
information from these sensors is sent to the
brain and enables us to see.
4
Color Theory
5
Color Theory
There are three types of cone. A somewhat simple
interpretation is that red cones are sensitive to
red light, green cones are sensitive to green
light, and blue cones are sensitive to blue
light. The perception of color depends on an
imbalance between the stimulation level of the
different cell types.
6
Color Theory
So when red light hits the eye, the cones
sensitive to red are excited, causing a
particular sensation in the brain. That is what
we call "red." This is what physically happens
when we observe a tomato in neutral daylight the
sunlight, composed of electromagnetic waves with
wavelengths from 380 nm to 780 nm, shines on the
tomato. The tomato absorbs the light from 380 nm
to 580 nm and reflects the light from 580 nm to
780 nm. This light reaches our eyes, where it
excites the cones that have their sensitivity
centered on 570 nm. The signals from these cones
are finally processed by our brain. This
particular sensation called "red", we are taught.
7
Color Theory
The relationship between wavelength and actually
hue is roughly 620-730 nm red590-610 nm
orange550-580 nm yellow490-540 nm
green450-480 nm blue380-440 nm violet
8
Color Theory
The colors we see are affected by the intensity
of light and by its spectral content. At low
levels of illumination, objects are less colorful
because rods are dominant. In bright daylight, we
see more color, contrast, and saturation because
cones dominate.
9
Color Theory
Additive color processes, such as television,
work by having the capability to generate an
image composed of red, green, and blue light.
Since the intensity information for each of the
three colors is preserved, the image color is
preserved as well. The spectral distribution of
the image will probably be wrong, but if the
degree of intensity for each of the primary
colors is correct, the image will appear to be
the right color. Red, green, and blue are the
additive primary colors, because they correspond
to the red, green, and blue cones in the eye.
10
Color Theory
11
Color Theory
Subtractive color processes work by blocking out
parts of the spectrum. The idea of subtractive
color is to reduce the amount of undesired color
reaching the eye. If, for example, you had a
yellow image, you would want to have a dye that
would let red and green reach the eye, and block
out blue. The additive secondaries become the
subtractive primaries, because each of the
additive secondaries will reflect two of the
additive primaries, and absorb one of the
additive primaries.
12
Color Theory
Additive Secondaries/Subtractive Primaries
Absorption Chart Color Reflects Absorbs
Yellow Red and Green Blue Magenta Red and
Blue Green Cyan Green and Blue Red
13
Color Theory
With this information, if we wanted red, we would
mix magenta and yellow. Magenta would absorb
green, and yellow would absorb blue, leaving only
red to be reflected back to the eye. For black, a
combination of all three would be used, which
should block out all light in theory. Printers
use black as well, since the dyes used in
printing are not perfect, and some light from
other parts of the spectrum gets through.
14
Color Theory
15
Color Theory
Sam Wilks would be annoyed, but colorful.
16
Color Theory
Color can be defined by three properties hue,
saturation, and lightness or brightness.
When we call an object "red," we are referring to
its hue. Hue is determined by the dominant
wavelength.
17
Color Theory
The saturation of a color ranges from neutral to
brilliant. The circle on the right is a more
vivid red than the circle on the left although
both have the same hue.
18
Color Theory
Lightness or brightness refers to the amount of
light the color reflects or transmits.
19
Color Theory
Color ordering systems, such as the Munsell
System, use the three properties of color to
identify unique colors. Notice that colors are
distributed in three dimensions.
20
Color Theory
We commonly see colors arrayed in two dimensions.
This is a useful, but incomplete representation.
Colors actually occupy a three-dimensional
space.
21
Color Theory
Lightness is the third dimension that is not
shown in color wheels often used in image editing
software.
22
Color Theory
To measure and predict the appearance of a
particular color, we need a way to link
perception to numbers and formulas. Scientific
color values were established earlier this
century by the CIE group. CIE models for defining
color space all rely on the same basic numbers.
23
Color Theory
CIE values were determined by testing human
observers. These perception-based values can be
used to measure and compare colors produced by
different methods of reproduction. The CIE
chromaticity diagram is one way to plot the
colors the human eye can see.
24
Color Theory
A color can be specified by its colorimetric
values. A colorimeter is an instrument that
measures color using numbers derived from CIE
values
25
Color Theory
A spectrophotometer is another instrument for
measuring color. It samples wavelengths across
the color spectrum.
26
Color Theory
Measuring color allows us to compare the color
gamut, or range of colors produced by different
methods. Color transparency film produces a wide
range of colors including some a monitor cannot
display.
27
Color Theory
Color printers and printing presses have
different color gamuts. They can never capture
all the colors that the eye can see or that are
in an original transparency, but they can
simulate the appearance very successfully if
color reproduction is understood and controlled.
28
Color Theory

• Hue  color with no black, white or gray added
• Tint  hue white
• Shade hue black
• Tone hue gray or hue varying degrees of its
  complementary color
• Value how light or dark a color appears
• Intensity how bright or dull a color appears,
  also called saturation and/or chromaticity.
  Basically, how much of the hue is identifiable.
  Grays are achromatic, meaning no hue/color.

29
Color Theory
An important characteristic of each component in
the color production system is its Gamma. It is a
number that indicates the relationship between
the signal values at input and output of a
particular device. A Gamma of 1 indicates a
linear behavior. This means that the device's
output is directly proportional to its input.
30
Color Theory
A color display may have a non-linear color
behavior. If the video signal at the input
contains a value of 60 of full-scale red, the
image on the screen may be only 30 of full-scale
red (thus dark-red). This is why a scanned image
normally looks darker when it is shown on a
display. This behavior can be graphically
displayed in a Gamma curve.
31
Color Theory
Range of Gamma Adjustments
32
Color Design
People with normal color vision are called
"trichromats" because they require three
primaries to match any arbitrary sample. The
trichromatic eye has three cone types, each
containing a photopigment which responds to a
restricted range of wavelengths. There are 7
types of color deficiency due to cone
abnormalities. In addition, the elderly see
colors differently, but are not color blind in
the usual sense of the term. Finally, brain
damage can create a very rare condition called
achromatopsia.
33
Color Design
Anomalous Trichromats There is a subpopulation
of trichromats, who still requires three
primaries to match a sample, but whose matches
are abnormal because they use one primary far
more than would be expected. While having all
three cone types, one cone type is rarer, has a
reduced amount of pigment, or has a pigment tuned
to an unusual wavelength.
34
Color Design
These "anomalous trichromats" fall into three
groups, based on the primary which they overuse
Type Cone Male Female Protoanomalous red
1.0 .02 Deuteranomalous green 4.9 .38 Tritana
nomalous blue 0 0
They can see all hue categories, so they are not
color blind in any real sense. But they may have
difficulty is discriminating colors, which a
normal would easily distinguish.
35
Color Design
Dichromats The second class of color abnormal is
the dichromat, a person who requires only two
primaries to match any sample. These people are
missing one of the three cones types
36
Color Design
Type Cone Male Female Protanope red
1.0 .02 Deuteranope green 1.1 .01 Tritanope
blue .002 .002
37
Color Design
Unlike color anomalous individuals, dichromats
are true color blinds in the sense that there are
some hues which they cannot perceive. Lights
which would appear different to a trichromatic
will be metamers (appear identical) if they
create the same activation ratio in their
remaining two cone classes. Protanopes and
deuteranopes are red-green color blind and see
only yellows and blues. The tritanope is
analogously blue-yellow color blind.
38
Color Design
Monochromats Monochromats can match any light
with a single primary. They generally have no
cones and make all matches using rods. They are
very rare, 1 in 10,000,000. With only a single
receptor type, they can have no color vision and
are truly color blind because they distinguish
only brightness levels. Their vision is so
generally poor that the color selection for
visual design is the least of their problems.
39
Color Design
Dichromats have a point on the spectrum called
the "neutral point" where the lights appears
achromatic. The point is about the same for both
classes, 495 nm for prontanopes and 500 nm for
deuteranopes, wavelengths which would appear
slightly bluish green to a normal. All
wavelengths below the neutral point appear blue
while those at longer wavelengths appear yellow.
In the first few nm. away from the neutral point,
apparent saturation increases rapidly, so
dichromats can easily discriminate colors. The
rest of the blues then appear similar as do the
rest of the yellows and only discrimination of
brightness is possible.
40
Color Design
Toward very long wavelengths, protanopes
experience lights as becoming darker, so a very
red apple, as already mentioned, might look dark
yellow or brown. Some people think that a
protanope sees red as black on a CRT because he
has no "red cones." This is false because CRT
colors always have white content and because
"green" cones are still slightly sensitive to
relatively long wavelengths. However, deep red
may appear dark. Tritanopes have a neutral point
at 570 nm., a wavelength which appears yellow to
a normal trichromatic. Higher wavelengths appear
red while lower ones appear green.
41
Color Design
The description of dichromatic color perception
comes from both theory and from studies of a few
people who were dichromatic only in one eye.
However, its not absolutely conclusive. First,
there is no guarantee that the "normal" eye of a
unilateral dichromatic is really normal. Second,
studies often find that dichromatic color vision
is much better than that predicted by theory. The
best guess is that rods activate the red
component red-green opponent process to give
dichromats a weak three-dimensional color space.
42
Color Design
The elderly see colors differently from the
young. In fact it is somewhat surprising that
designers worry so much about color blindness,
which is 1 (dichromatic) of the population,
while they often ignore the much larger group of
elderly, who almost always exhibit visual
deficits. The elderly (65) are 12 of the
population and those just beginning to experience
visual decline (50) are 25.
43
Color Design
Vision declines with age in several ways, but the
most relevant for color design is the yellowing
and darkening of the lens and cornea and the
shrinking pupil size. Yellowing selectively
blocks short wavelength light, so blues look
darker. Moreover, the elderly have difficulty
discriminating colors which differ primarily in
their blue content blue-white, blue-gray,
green-blue green, red-purple, etc.
44
Color Design
Aging also reduces the amount of light reaching
the photoreceptors compared to the young viewer.
All colors will be dimmer and visual resolution
lower. For example, a moderately bright yellow
may appear brownish and dimmer blues will appear
black. When designing for the elderly, use bright
colors and make sure that brightness contrast is
especially high (and text larger) to help
compensate for acuity loss.
45
Color Design
Principal use of color Conveying emotion and
meaning Changing perception of space Changing
apparent size Showing similarities and
differences Parsing the visual field into
chunks Linking spatially separated objects
together Attracting attention Creating
emphasis Smoothing to improve image quality
Creating aesthetic and emotional appeal
46
Color Design

• Various sources have suggested other standard
  meanings signaled by color
• red urgency, passion, heat, love, blood
• purple wealth, royalty, sophistication,
  intelligence
• blue truth, dignity, power, coolness,
  melancholy, heaviness
• black death, rebellion, strength, evil
• white purity, cleanliness, lightness,
  emptiness
• yellow warmth, cowardice, brightness
• green nature, health, cheerfulness,
  environment, money, vegetation

47
Color Design
Color can affect perception of 3D space.
Mountains off in the distance usually appear
blue-violet and indistinct. This effect, "aerial
perspective," is a secondary consequence of short
wavelength lights greater refractive index - it
bends more. When white sunlight hits the
atmosphere, the short wavelengths refract more,
scattering into a blue haze - hence a blue sky.
In general, the further light travels from an
object to the eye, the more blue haze you see
through and the less sharp the object appears.
48
Color Design
The eye automatically interprets blue/violetness
and loss of sharpness as signs of distance.
Conversely "warm" colors such as red, yellow and
orange appear closer. A clever designer can use
these color properties to add a 3D feeling to a
flat display and enhance separation of foreground
and background.
49
Color Design
The foreground-background separation works best
when foreground color is bright and highly
saturated while the background is desaturated.
Notice how the text and rectangles seem to float
above the background on the violet side but not
on the red side. In fact, note how the entire red
side appears to be covering the violet, so that
the violet squares appear to be "holes" in the
red which reveal the violet underneath.
50
Color Design
Hue is good for showing categorical (nominal
scale) distinctions. People naturally break the
hues into categories, which they quickly and
easily judge as being the same or different. This
makes color an ideal way to signal that objects
have similar or different meaning, function,
importance, etc. One can easily distinguish 12 or
so colors red, green, blue, yellow, orange,
purple, pink, brown, cyan, black, gray, white.
51
Color Design
Hue is poor for showing quantitative
relationships, more/less, bigger/smaller, etc.
because there is no natural ordering. Some
advocate use of the "rainbow scale,"
red-orange-yellow-green-blue-violet, as a natural
ordered representation with blue as less and red
as more. People may have learned this order to an
extent, but it is not a very powerful innate
perception. If it works at all, its probably
because it approximates a brightness scale blue
is dark, green is middle brightness and yellow is
high brightness.
52
Color Design
A good design will use color to save the viewer
from thinking too much. With several hundred
million years of evolution behind us, color
perception is deeply wired into our fabric. As a
result, color perception is fast, accurate,
automatic, and effortless. On the other hand,
thinking is a relatively recent evolutionary
advance, and we are not yet very good at it.
Thinking means reading text, attaching meaning to
an icon, searching memory, etc. These activities
are relatively slow, error-prone, require mental
resources and effort and take learning. How many
times have you misread a word? Probably often.
How many times have mistakenly seen blue as
orange? Probably never, even if you are a
dichromat. In sum, a designer should strive to
capitalize on our hard-wired perceptual
capabilities whenever possible.
53
Color Design
Similar colors suggest a similarity relationship
and different colors a difference relationship
between objects and areas of the screen. It could
be that names or icons representing all image
files are red and all text files are in green or
that last years sales are blue. At first, the
viewer wont know that colors refer to different
files or sales classes, but he will automatically
know that similar color represents similar type.
The formation of the association through color is
wired into us, so it is very fast, efficient and
effortless. More importantly, it has no capacity
limitation so we can see these relationships
across the entire visual field at a glance.
54
Color Design
Objects of different color are automatically
divided into different categories. It is very
hard to attentionally group together objects of
different color. Heres a simple example. Compare
the effort required in reading the two words.
Color variation makes grouping the letters into a
readable word much more difficult.
55
Color Design
The picture below demonstrates both preattentive
pop out and linking. Note that the red dots
immediately pop out from the blue and that you
can't help but link the red dots into a single
line. On the left, the dots differ only by
brightness, and the effect is not as strong.
Moreover, look at the word "pop out" in the
sentence above. In your peripheral vision, the
red still pops out while the brighter dots do not.
56
Color Design
Color in Animation The chromatic system has not
only poor spatial resolution, but also very poor
temporal resolution, a fact which can be put to
good use in making animations appear smoother
with fewer frames. Animations create the
sensation of motion by showing a succession of
still images. If done improperly, two artifacts
appear, each of which is improved by substituting
color contrast for brightness contrast.
57
Color Design
Flicker and jerky motion. High contrast
achromatic borders make motion appearance jerky.
Motion will appear smoother with low achromatic
contrast, but of course this makes the moving
objects less distinct. Again, the solution is to
simultaneously reduce brightness contrast and
increase color contrast. This effectively creates
motion blur by taking advantage of the chromatic
systems lower resolution. As with antialiasing,
the substitution of chromatic contrast for
brightness contrast "smoothes" perception and
greatly reduces flicker and jerkiness.
58
Color Design
Multiple images. Sometimes, as objects move
rapidly across the screen, the viewer will see
multiple versions of the image at neighboring
spatial locations. This occurs because the high
spatial frequencies, in which we see fine detail,
have longer visual persistence. The chromatic
system's lower spatial resolution means that it
will not see the higher frequency components, so
again, the designer can substitute color contrast
for luminance contrast.
59
Color Design
Color Schemes Monotone achromatic neutral colors
only, white, black gray. Easy to use, but
requires expert care to avoid complete
boredom. Monotone chromatic a single chromatic
color ranging in brightness and saturation red,
pink, rose, etc. Easy to use, but again risks
monotony. Analogous hues two or three colors
close to each other in color space blue-greens,
etc. Easy to use.
60
Color Design
Complementary hues contrasting hues
blue-orange, red-green. Requires care since it
can appear garish. Works best when the two differ
significantly in brightness and one is relatively
desaturated, e.g., a dimmer, desaturated red and
a brighter saturated green. Split complementary
hues uses three colors, one base color plus two,
near, but not directly across, color space
orange with blue-greens. Works best when colors
from opposite sides differ significantly in
brightness and color(s) from one side is/are
desaturated. Usually produces a less garish
feeling than a simple complementary scheme.
61
Color Design
Triad hues uses three hues approximately
equidistant in color space red-yellow-blue.
Desaturate at least two hues. Triad schemes are
often embedded in a generally neutral field, so
it would work well on a gray background. Tetrad
color uses 4 hues a equal intervals in color
space yellow-green-orange/red-blue/violet. Leave
this to the experts.
62
Color Design
Secondary
Primary
Tertiary
63
Color Design
64
Color Design
Analogous
Complementary
Essentially Complementary
65
Color Design
Some common mistakes in color usage Using
insufficient brightness contrast. This is the
biggest sin, by far. Paying attention only to
aesthetics. Using color for colors sake without
a specific plan or color scheme. Assigning
different colors to the same type or the same
color to different types. Using hue to represent
a quantitative continuum. Creating large fields
of saturated color. Using too many colors.
66
Color Design
Web Color Design These are the 216 colors that
the Netscape browser uses in its palette. By
using these colors on your web pages, you can be
assured that the viewer will see your images
exactly as you intended. If you don't use the
color cube, Netscape will use it for you.
67
Graphic Design
68
Graphic Design
69
Graphic Design
70
Graphic Design
71
Graphic Design
72
Graphic Design
73
Graphic Design
74
Graphic Design
75
Graphic Design
76
Graphic Design
77
Graphic Design
78
Graphic Design
79
Graphic Design
80
Graphic Design
81
Graphic Design
82
Graphic Design
83
Graphic Design
84
Graphic Design
85
Graphic Design
86
Resources
Color Theory http//www.beer.org/tpark/color.html
http//www.kodak.com/US/en/digital/dlc/book3/chap
ter2/index.shtml Color Design http//www.ergogero.
com/FAQ/cfaqhome.html http//www.killersites.com/1
-design/ Graphic Design http//www.mundidesign.com
/presentation/index2F.html http//www.graphicdesig
nbasics.com/