Transcript Theory of Vision: What We Can Easily See

Text Lecture 2
Schwartz

The problem for the designer is to ensure all
visual queries can be effectively and rapidly
served.


Semantically meaningful graphic objects need to
have the right amount of salience.
Visual queries should be supported with the most
visually distinct objects.



The PVC has been mapped
 by Hubel & Wiesel, (1981)
The PVC has a number of
interlocking layered sub-regions,
each of which respond to either
color, shape, texture, motion, or
depth.
The PVC sub-regions are– V1,
V2, V3 V4, and V5
http://en.wikipedia.org/wiki/Visual_cortex
1 million nerve
fibers for each eye
Billions of neurons in the PVC
Working together in PDP
(parallel distributed processing)
The PVC has a number
of interlocking layered
sub-regions, each of
which respond to either
color, shape, texture,
motion, or depth.

V1 and V2 provide inputs to two
distinct processing systems.

The what system

Responsible for
identifying objects in
the environment.

Helps identify if a
pattern of light and
color represent an
identifiable object.

The where system

Responsible for
determining the
location of information
in the world and
guiding actions to grab
it, or navigate around it
or to it.
What and Where Pathways
What pathway sweeps forward from V1
and V2 along the low edge of both sides
of the brain.
Where pathway sweeps higher up on the
brain and processes information about
where objects in the world are located.




Perceptual attention is called
pre-attentive processing, but
it is all about tuning. And
most of it occurs in the
sensory register.
Tuning refers to the
cognitive processes
mobilized from the visual
properties of graphics
(salience) used to plan the
next eye movement.
In pedestrian terms: “pop
out”
“Pop out” requires a 30
degree orientation
difference to stand out.
Effects
Due to feature-level CONTRAST.
SIZE
MOTION
COLOR
ORIETATION
CONTOUR (CONVEX/CONCAVE)
Creates visual distinctness
Requires only ONE eye fixation.
Perceived in 100 milliseconds

Trying to find a target
based on two features






Size & color
Orientation & contour
Etc.
Most visual conjunctions
are hard to see.
Conjunctions require
attention; non-conjunctions
do not.
Time to do a conjunctive
search increases linearly
with the number of items.
 You have to pay
attention to each item,
one item at a time to
bind the features.



Neurons sensitive to more
complex conjunction patterns
are found farther up the
“What” processing pathway.

These neurons cannot be
used to plan eye
movements.
If background is very
homogenous, then a small
difference is needed to make a
feature distinct.
The more the background
varies in a particular feature
channel (color, texture,
orientation), then the larger
the difference required to
make a feature distinct.






Color
Size
Shape
Orientation
Motion
Contour





To make something easy to find, make it different from
its surroundings using a primary visual channel.
To make several things easily searchable at the same
time, use different channels.
More than one feature can be changed on the same
channel.
Using more than 2 or 3 symbols to create “pop-out” is
difficult.
When aiming for pop-out, there are only three
difference steps on each channel: 3 size, 3 orientation, 3
frequencies of motion, etc.


What takes a fraction of a second with multi-feature
designs can take several seconds with a consistent (say,
color-only) design.
Making visible enhancements is better than visible
reductions.






Increase symbol size rather than decrease it.
Increase contrast rather than decrease it.
Add extra parts to a symbol rather than take parts away.
Manipulate motion in the periphery of the visual field.
Motion should not be “shaking” per se, rather it should
move into the visual field.
Fast motion is annoying; slower motion is alerting



The goal is to always to get a visual target in the
vicinity of the eye’s detection field. From there, it
becomes eligible for the next fixation.
The bigger something is, the more it takes up in the
detection field.
Salience of visual channel targets create:
HIERARCHICAL SEARCH STRATEGY.
This poster shows multiple feature channels and visual searches. This duo tone
poster has color as the top of the visual hierarchy. The blue and yellow have a pop
effect, bold, bright colors on a black and white poster. These color features are
also arrows that bring direction and motion to the poster. This keeps the eyes
moving back an forth across the page. The bold black text boxes keep your eyes
centered on the page. The dark irregular shape of the monkey then attracts the
viewer. Color, motion, then shape. POSTED BY BAILEY KIMA
http://f09323bnakano.blogspot.com/

Move and Scan Loop


Eye Movement Control Loop


Is the first step: involves orienting the head and
trying to get the best viewpoint.
Is the second step: involves executing eye
movements around salient features and feature
categories to map the locations in the whole design.
Pattern Testing Loop

Is the third step: involves finding a promising target
and testing the pattern to see if it is the search target
or not.