Low Vision and the Visual Interface for Interactive
Television
Mark Rice
Deborah Fels
University of Brighton
United Kingdom
Ryerson University
Information Technology Management
350 Victoria St. Toronto
Canada
M.D.Rice@bton.ac.uk
dfels@ryerson.ca
Faculty of Information Technology
Lewes Road, BN2 4BR
Abstract
Viewers with vision impairments often have difficulty seeing all of the content on television and now
they have the added task of seeing and interpreting interactive elements that are primarily image or
graphics based. We carried out two studies to examine the difficulties partially sighted and blind
users have with current interface designs for interactive television. Results indicate that people’s
visual acuity, contrast sensitivity and colour perception differences affected their ability to
comprehend the displayed information making it difficult to have a standard set of interface designs
that would accommodate the large variety of needs. A series of guidelines to accommodate such a
heterogeneous user group are thus provided.
Keywords
Interactive television, visual impairment, accessibility, visual design, interactive elements.
Barriers of Interactive Television
While the development of digital interactive
television has the potential to improve
television viewing, many issues as well as
unique opportunities for users with reduced
physical and cognitive disabilities, such as the
elderly and visually impaired arise. In the UK,
people with vision impairments represent a
substantial viewing audience, with 94%
watching TV on a regular basis [9]. Yet poor
interactive executions in the user centered
experience, negative pre-conceptions and
comfort with existing television viewing
patterns presents barriers to iTV access for
people with low vision [8]. The feasibility of this
medium
in
providing
greater
social
independence through public and education
services are being challenged because of the
difficulties caused by lack of appropriate user
interface modalities (e.g., audio description, text
and enhanced closed captioning or subtitles,
and customizable properties of interactive
elements) that suit a highly diverse and
heterogeneous viewer population.
An important factor in the development of iTV
is the lack of adaptive technologies, or
compensatory tools that can support people
with visual, auditory, cognitive and manual
disabilities in the acquisition of interactive
television information. In the computing world,
there are many available adaptive technologies
such as screen readers, voice recognition
systems, and adaptive handheld/keyboard
device for people with vision impairments [1].
However, far fewer adaptive technologies are
available for television likely because analogue
television technology is so limited. For example,
adaptive technologies for people with vision
impairments are limited to a series of lens
systems, which include hand-held stand
magnifiers, telescopic lenses and front screen
magnifying displays and tactile indicators for
remote control devices. Presently, interactive
television uses a high level of visual elements
assuming that users have significant visual
competence to be able to recognize the graphical
information presented on-screen. If they do not,
there is little available to support the need for
alternative methods
information.
of
presenting
this
There has been some research carried out to
determine visual properties of computing
displays and preferences for people with vision
impairments [3, 11,]. The results of these studies
provide design recommendations to improve
navigational access within computer interfaces.
However, they were carried out with high
resolution screens in productivity or searching
types of tasks.
In this paper, we will provide research that
attempts to analyze defining factors in image
size and icon presentation that influences
successful interaction for visually impaired
users of interactive television.
Goals of the Research
The first study in our research had three goal: 1)
to determine the ease or difficulty of registering
graphically based information on digital
television screens, through a range of sight
conditions; 2) to understand the problem
solving techniques used by people with vision
impairments to determine what is displayed;
and 3) to evaluate the optimal quantity and size
of interactive elements and information,
recommending accessibility guidelines for
content designers. Two studies using a common
set of participants and room design that
addressed these goals are reported.
Participants
Twelve participants were recruited with the
help of local disability organizations within the
East Sussex area (UK). These participants took
part in two separate studies designed to address
our research goals. Ages ranged from 19 to 87,
with an average age of 38 years old. There were
five female and eight male participants. This
may not be truly representative of the national
visual impaired age, because approximately
three quarters of all visually impaired people
are over the age of retirement [3]. Table 1
provides a distribution of vision impairments of
the studied participants1. All participants were
encouraged to bring a sighted friend or relative
for support, although only three people did this.
In addition, two people also had hearing
impairments and one person experienced
dexterity difficulties using the handset control.
Following consent to participate, users were
asked questions about their television viewing
habits, visual disabilities and accessibility
restrictions in using television in an interview
form.
Diagnosis
Number of Subjects
Glaucoma
1
Diabetic Retinopathy
4 (one had cataracts)
Macular Degeneration
(MD)
2 (one was blind in left
eye)
Retinoblastoma
1
Nystagmus + Cataracts
2 (one had cataracts)
Congenital Rubella
1
Lebers Optic Atrophy
1
Table 1: Distribution of vision impairments of
participants.
Room Design
Two studies were set up using the facilities in
Brighton University’s usability laboratory. The
laboratory consists of a 4 X 3 meter room, fitted
out as a domestic lounge or living room. It
contains two sofas, a coffee table, standard lamp
and a wide screen domestic television. Set in a
side wall is a two-way mirror that enabled the
occupants of the room to be observed without
distraction. Close-circuit television cameras on
motorized mounts within the room enabled
recording of activities by the occupants. The
television display could also be recorded
directly. A microphone in the room picked up
conversation.
The content displays were created as HTML
web pages and broadcast through a Bush
Internet box. This set-top box allows web pages
to be accessed from the television using an
ASDL telephone line.
Information was displayed on a 26” television
monitor, using a 16:9 display format. In two
cases, participants requested a 4:3 format, as it
provided them with a better screen resolution
size. To prevent unnecessary glare and light
reflection on-screen, the room was dimly light,
with one overhead 60 watt lamp placed 0.7
meters above the sofas.
During testing, the recorded videos were
formatted in MPEG, and transferred to CDROM for further evaluation.
Study 1: Icons (size, color and shape)
The purpose of this study was to identify the
visual properties of icons (color, size, and
location) necessary to make them legible on
television for visually impaired audiences. This
study is an adaptation from a previously
published experiment that examined the visual
performance of low sighted users’ ability to
recognize graphical elements within a
traditional windows desktop environment [3].
The entire study procedure was verbally
explained to all participants in the experiment
in advance. For training on the study protocol,
users were asked to recognize a stimulus icon
that was randomly displayed on the television
monitor for two seconds. The size of the icon
was large (180 mm x 120 mm) as recommended
by Jacko et. al., [3] to maximize the probability
of detection. Users’ tasks where to match the
original stimulus icon to a target icon, on a
second presentation screen, by verbally
indicating its position by number, from a
maximum of six horizontal positioned icons (far
left: 1, to far right: 6)
For the study, a sample of six different icons
was taken from present designs in interactive
services in the UK (Sky, BBC Freeview), in order
to replicate a realistic user experience (see Table
2).
24, and 54 x 36 mm). All the icons on the single
screen were the same size, through randomly
displayed. The coloured background on which
the icons were presented was also manipulated
from five different saturated colours - black
(#000000), white (#FFFFFF), red (#CC0000), blue
(#000066) and green (# 39900).
Viewing distance to screen, icon size preference
and reaction time, recognizing the original
stimulus on-screen and then verbally locating
the target icon from the stimulus were recorded.
Noldus behavioral software was used to
determine the exact reaction time from the
video and audio recordings.
Results
Recognition of icons varied largely depending
on the number, shape, color and size of each
icon displayed. No visual cues or prompts were
used during this experiment. Two people did
indicate that they might use a visual aid, such as
a handheld screen magnifier to assist in the
access of similar sized icon sets on interactive
services. Table 3 presents the mean reaction
times for the various icon types selected by the
participants. In all instances, the smaller size
resulted in slower reaction times.
Icon type
1
2
3
4
5
6
25 x 16
6.8
8.8
9.7
9.8
10.2
7.1
36 x 24
6.3
8.7
9.5
9.2
9.9
6.3
54 x 36
5.7
8.3
7.9
8.9
8.7
5.4
Table 3. Mean reaction times per icon type
Icon
type
1
2
3
4
5
6
Table 2: The six icon types used in study 1.
Using a 3 x 5 x 5 repeated factorial design three
independent variables were examined - icon
size (3), background colour (5) and set size (5),
totaling 75 conditions. Users were asked to
select an icon from the set displayed on the
target presentation screen and match the
stimulus icon presented on the first screen.
Following the same design strategy, sets of two
to six icons were systematically presented on
the target screen. On each trial, the target icons
varied over three different sizes, (25 x 16, 36 x
1.Shape
A defining characteristic for visual acuity is the
shape of the icon [4]. Nine (88.7%) of the
participants indicated that shape was important
to them in icon recognition although icons were
repetitively misunderstood.
Of the six icon shapes presented during this
experiment, the two axed, diamond-shaped
icons (colored blue and yellow) seen in Fig. 1
caused the greatest visual comprehension
difficulties.
as the stimulus icon (individually) in the five set
than any other set size. When they were both
paired in the target screen they appeared almost
identical, and became very difficult to
differentiate between.
Figure 1: Horizontal and vertical chevron
icons.
People indicated that the 2 to 12 mm
(depending on the icon size) space or line in the
center separating the directional arrows in the
blue and yellow icons was too narrow making
the chevrons look fused together as one shape.
The result was that both icons looked identical
as seen in Fig. 2. This problem was amplified as
the icon size decreased.
The results illustrate the importance in
providing enough distinction between opposing
shapes, through appropriate color difference
and spatial layout, if understanding functional
meaning is to be afforded by icons on television,
particularly when these icons appear in sets of
icons as there must be enough distance
separating them to make them distinguishable.
However, the optimal distance between
opposing icons or between unique icons
remains to be determined and will depend on
the display size intended for that icon.
Figure 2: Horizontal and vertical chevron
icons.
Participants had to expend much perceptual
effort on trying to distinguish the differences
and meaning of these two icons as illustrated by
the following participant comments:
“It’s such a thin line between the two triangles, you
have to look harder to see what way that line is
going”.
“The diamond was harder because it was more to
look at...without a struggle I can’t see whether I have
to go left right, up or down”
“I find that one very hard to distinguish if I was not
concentrating on the vertical, horizontal split”.
The understanding of the fused diamond icon
was largely unknown by most participants. Few
recognized that the icons used might represent
directional arrows. When one or more of the
directional arrow icon designs was used within
icon sets with two or more elements, recognition
and separation became particularly confusing
for people with low vision. The recognition
difficulty translated into increased reaction
times as seen in Fig. 3. For example, in five icon
set, two similar style icons presented (type 3 and
4) were the most difficult to interpret, mainly
because overall they were displayed more times
Time frequency (sec.)
12
10
8
6
4
2
0
2
3
4
5
6
Icon set
type 1
type 2
type 3
type 4
type 5
type 6
Figure 3: Mean reaction times per icon set. (See
table 1 for definitions of icon type.)
2. Color
In addition to shape, color highly influenced the
legibility of the icons presented. Figure 4 shows
the color combinations used in this study along
with the mean reaction times. There was a
significant difference in the mean reaction times
between using white (11.0), green (5.9), blue
(7.0.), black (6.1) and red (7.9) backgrounds.
High luminance background colors (such as
white at 100% luminance) were reported to
cause excess glare making it difficult for people
to see. Participants also had difficulties with red
(luminance 30%) and reported icons with red in
them as being too bright. This caused similar
inference to the visibility of graphical
information on-screen as white. As seen in Table
4, the fastest reaction times in finding the target
icon were found with colors that gave strong
overall contrast to the icons selected (e.g. black
and green).
Icon recognition was further complicated for
people with congenital color defects, as
indicated by the following participant
comments:
“It looks like a black cross, with a white background,
or a funny puzzle, or it’s two chevrons”.
“That blue virtually turns to black when you look at
a white screen”.
With foreground colour combinations, not all
colours were equally distinguishable. Using a
parallel analysis of Murch’s [6] colour
guidelines, a series of comparisons were found,
in choosing foreground and background
combinations. The two most striking similarities
included:
 The careful grouping of screen elements, so
that background colours do not change the
element in the group. With evidence that lighter
backgrounds make elements smaller and
darker, while darker backgrounds make
elements appear larger and lighter.
 The avoidance of red in the periphery
display, as the retinal periphery is less sensitive
to detect this colour.
Time frequency (sec.)
25
20
15
10
5
0
black
white
blue
green
red
Screen background colour
Mean
Max duration
Min duration
included white on black, yellow on blue and
yellow on purple.
Differences in color perception attributable to
abnormalities
in
color
vision
effects
approximately 10% of males and 0.5% of the
females [2]. It is important and simple from a
design perspective to alter color combinations
used in icon design to address this difficulty [4].
As mentioned above, background contrast made
the same icons on different screens appear
lighter or darker. In particular, dark
backgrounds, such as black increased the image
colorfulness, while white, red and blue
appeared to reduce it.
Relying on color only to convey representational
meaning should also be avoided. While
appropriate color consistency and contrast
differentiation between icon sets can aid in the
visual recognition of the icons, content
designers should ensure that icons have enough
contrast or lightness to separate them from the
background television display.
3. Size
Overall, 25% of participants could not recognise
sufficient detail in the smallest icon size to
distinguish the target. Reaction times did
improve with larger sizes (>36 x 24 mm).
Nevertheless, 16.6 % of users struggled to locate
the appropriate icon even with in the largest
icon size (54 x 36 mm). During the study each
person were tested using all the three target
sizes. While there is evidence that larger sizes
improve icon legibility and reduce reaction
times (See table 3), variations in screen
proximity during the study means that further
research is necessary to verify the effects of size
on peripheral and central vision loss.
4. Speed of display
Three participants found that the stimulus icon
was too briefly displayed on-screen. Overall, for
the user group to recognize the stimulus from
the target icon a mean time of 6.7 seconds was
recorded. Longer reaction times were caused by
physical eye strain, attributed to the repetitive
nature of the experiment.
Figure 4: Mean reaction times per background
color.
Study 2: Images
Selection of foreground colour is highly
dependable on the background colour. The
most legible combinations in this study
appeared to be those with good contrast, these
Study 1 showed the importance of shape, and
color for icon design and acquisition for people
with low vision. However, we also wanted to
determine the effect of size and icon design on
viewers’ ability to understand the purpose of
each icon. The purpose of the second
experiment was to examine the effect of size,
contrast and image representation on the visual
comprehension of viewers with low vision.
each image and the physical tiredness the
observer.
Design
Participants were first presented with fifteen
numbers displayed on-screen and instructed to
select one. Each number revealed a JPEG image
selected from three common scenes, landscape,
portrait or simply-shaped objects (flowers, balls,
sign posts), each displayed in the top right-hand
side of the screen, sized 38 x 68 mm. After
selecting the number, participants were
instructed to verbally describe the image (i.e.,
what they could recognize and distinguish in
the picture). After this initial description, the
image size was doubled (i.e., from 38 x 68 mm
to 76 x 134 mm) (see Figure 4 for an example).
The participants were then asked to elaborate
on the description of the image. This process
was repeated by enlarging the image seven
times, each time increasing the linear size by 38
x 68 mm, as seen in Table 4.
Image Display
1
2
3
4
5
6
7
Size (mm)
38 x 68
76 x 134
114 x 200
152 x 266
190 x 332
228 x 398
266 x 464
Table 4: Progressive images sizes viewed by
participants.
During this study, participants were asked to
describe what they could recognize, were
uncertain
about,
or
had
difficulty
understanding. They were prompted with
questions from the researcher on the pictorial
content, and asked to indicate when they
thought the image was large enough to
completely comprehend its contextual meaning.
For example, if the image was a portrait of two
people, the researcher would ask questions
regarding the sex and age of people, to describe
appearances, location, and actions (i.e. what the
people in the portrait were doing). Four to eight
different images were used per trial. The
number of images participants described was
dependent on the amount of time taken to view
38 x 68
152 x 266
266 X 464
Figure 5: Illustration of the image enlargement
process.
Results
Interpretation of image meaning
A preliminary review by one independent rater
on a general scale of closeness to an original
description of the content was carried out,
evaluating several important categories, e.g.,
description of people, scenery and general
attention to detail.
A series of striking
observations were made between size,
familiarity, movement, distortion and contrast
of the images displayed.
1. Image size
For 11 (91.6 %) of the participants, the first size
(38 x 68 mm) was very poorly interpreted. Often
participants could only recognize two or three
color combinations, or singular shaped objects
such as rectangles or squares. Comprehension
improved with larger image sizes (>190 x 332
mm), particularly with higher image resolution,
as detail became more apparent. Not surprising,
the best image size was 266 x 464 mm.
Nevertheless, even at the largest size there was
still considerable variation between the
interpreted meanings of the images and the
intended meaning. It also became apparent that
many of the variations were due to higher
visual processing demands amongst subjects.
Greater sensory and information processing
demands were apparent for observers with low
visual acuity. Even with the display of a full
screen image (266 x 464 mm), substantial detail
was misinterpreted or left out of the verbal
analysis
(e.g.,
recognizing
a
persons
approximate age or sex). Perceptual difficulties
included reading text (dependent on size);
identifying objects (balls, clothes, and trees) and
information processing difficulties included
understanding context cues. As the image was
enlarged, participants were able to provide
descriptions with increasing amounts of detail.
However, 10 people (83.3%) were still unable to
provide complete details. This is because images
appeared fragmented along many dimensions
(e.g. color, orientation, motion factors) affecting
participants’ ability to decipher the number of
picture elements and degree of detail on screen.
Furthermore, picture processing is dependent
on a variety of conditions, requiring both
environmental and contextual knowledge of the
user to render a complex understanding of
image learning [10].
Clearly there were too few subjects for each type
to diagnosis an optimum image size. The
difficulty in determining the optimal image size
remains dependent on many design factors. For
example, in some cases, doubling the image size
did not necessary help to verify a clearer
interpretation of the image, in other instances,
substantial enlargement significantly aided in
bringing greater contextual understanding.
Variations in attention span, and participants’
ability to comprehend pictorial information, in
relation to their visual profile weighed heavily
on their perceptual understanding. Common
sitting patterns emerged, where viewers were
observed to physically move around the
television screen, tracing the compositional
arrangement of the image displayed.
Although a finer grain analysis remains to be
carried out to understand what specific parts of
the content cause the greatest difficulties to the
most number of people with vision
impairments, it has been found that image size
is not the only influential screen variable.
2. Familiarity and recognition
During the study, one participant with
Nystagmus (a condition that causes involuntary
and oscillated movement of the eye) explained
how he was able to interpret what he described
as a ‘green blob’ in the background of the
picture as grass. He did this by deducing from
what was visible in the foreground of the image.
Recognizing that there was a car, a fence and a
series of trees, he logically deduced that the
green shape had to be grass. Similarly, another
observer understood that the person was likely
to be a male, simply because he was wearing a
tie. Awareness to the characteristics of
surrounding or relative objects, from the
information available, helped decipher the
missing information. In both cases, participants
were able to extrapolate from recognized
contextual cues to help interpret something that
was either too small or too obscure to
comprehend without image-enhancing, low
vision aids. The degree of extrapolation varied
depending on the certainty of what was
understood in the picture. As one participant
indicated:
“You hypothesize when you are trying to recognize
something, and then you look for confirmation of the
hypothesis”.
The amount of image searching required also
had an impact on the search time required to
discover the image’s meaning. Participants with
central and peripheral vision loss took
substantially longer deducing an image than is
expected from a normally sighted person.
Familiarity with the image seems to help reduce
reaction time, particularly if it is something the
person is used to seeing.
“Once you know that its water, then it as case of
what sort of water is it. What kind of beach is it?
What kind of shore is it? Most things I get tend to be
big, broad brush strokes. Once you’ve worked out
what the picture is, it’s a case of pulling out the
detail”.
For others too much detail created confusion
and ambiguity in visual representation. Audio
therefore becomes fundamental in depicting
and interpreting the pictorial context on-screen.
In addition, 33.3% of the group reported a
preference for subdued or dimmed lighting to
avoid glare and excess light. Participants
reported preferring the cinema to home viewing
because of the large screen size and darkness of
the space. Low light conditions reduce
difficulties of glare and excess light as well as
high environmental contrast.
3. Movement
Two observers commented that movement
helped in the comprehension of detail on screen,
particularly in recognising people. Motion drew
attention to prominent visual characteristics,
such as colour and shape, which could be
tracked in ‘short bursts’, (depending on velocity
and size). A prime example of this was one
observer who was able to recognise the red
shirts worn by a football team. By tracking the
clusters of red shapes, he was able to gain some
valuable insight into the direction of the ball.
4. Distortion
Blurring or defocusing of an image can distort
parts of the picture, causing confusion and
uncertainty. For low sighted participants this
caused an additional level of unnecessary
abstraction and ambiguity.
5. Contrast
Objects with similar hue combinations did cause
differentiation difficulties, particular for people
with central field defects. Many people with low
vision have a reduction in contrast sensitivity.
Previous psychophysical research has reported
that contrast reduction effects people with both
central and peripheral vision loss in the
performance of reading tasks [5]. Contrast
discrimination in image representation has been
reported to be effected by luminance and
variations in spatial frequency [7]. While there is
evidence that adjacent objects with similar hue
levels add obscurity, or lose of information, high
contrast, particularly between the foreground
and background display may help to emphasis
the shape of an object, but also reduce the ability
of some individuals to see a high level of detail
in the image.
6. Screen proximity and physical movement
Common sitting patterns emerged, where
viewers were observed to physically move
around the television screen, tracing the
arrangement of the image displayed1. This
affected their search strategy and time in
determining an image’s meaning. Participants
with central and peripheral vision loss took
substantially longer deducing an image than is
expected from a normally sighted person mostly
due to the amount of searching and moving a
person required. As seen in Figure 6, 58.3% sat
within approximate distance of 30 cm of the
screen (measured from the eye).
Close proximity to the screen was reported as a
normal occurrence, and depending on the
activity, users often had a designated chair or
beanbag allocated next to the television screen.
For many participants, televisions and computer
monitors at home were often situated at a height
that was far more accessible for them to view
content:
“Both my television and computer screen are not this
wide, and my computer screen I have right in front of
me, close over the keyboard, and therefore it does not
take more movement then the head then that
(indicating a slight jolt to the head)”.
Participants favored images to be displayed
higher on screen because it meant less head or
body movement that may cause strain problems
such as back strain from bending in front of the
screen for too long. Close proximity to screen
reduced the field of vision, causing more intense
attention to the screen. Longer reaction times
arose from fixation problems, and higher
cognitive workloads caused by the small
amount of on-screen information seen at any
one time. Viewing a small section of the screen
at once raises questions to the viability of
understanding the relationship between objects,
and objects that are location sensitive, such as
the menu bars and navigational menus. To
compensate for this, two users’ suggested
placing visual information within a centralized
screen area.
Figure 6: Sample screen shot of a participant’s
close proximity to screen.
Discussion
From the research results, it became evident
that there are significant limitations in the
access of interactive television for vision
impaired users. Realistically, optimizing
parameters for a single screen user are
potentially problematic due to the high
number of variables and user requirements
that can considerably affect screen interaction,
and customized interface design strategies are
still being debated. Furthermore, interaction
1
Please note that distance to screen was not
controlled factor during the study, primarily due to
the range of differences in visual acuity.
between multiple factors must be accounted
for, such as proximity, colour and shape of
graphical elements displayed. Nevertheless,
taking this into consideration, there are some
general design guidelines that can enhance the
usability of interactive services for wider
audience groups:
 Color: Colored icons should be placed on
a neutral or desaturated colored background
in order to ensure maximum legibility.
Recommended combinations, yellow on blue,
white on black, yellow on purple. Exaggerate
lightness between the foreground and
background display. Avoid pairing red, green,
and blue (difficult to register for color blind
users).
 Shape: Keep the shape and design of the
icon simple and yet functional. Question
whether the shapes are distinguishable in icon
sets and if differences between icons in sets
can be distinguished even in smaller sizes.
 Position: Ensure icons are consistently
positioned throughout the screen design.
Avoid placing them directly next to text
dialogue boxes, and menus to prevent them
from being obscured, or partially hidden.
Provide sufficient spacing between each icon.
 Visibility: Bright backgrounds increase
the likelihood of screen flicker. Consider the
use of highlighting to draw attention to active
elements, such as icons on screen.
 Consistency
mapping:
Reduce
the
cognitive load of the user. Make sure
graphical elements are consistently positioned
in the same location across screen pages,
ensuring their functionality remains the same.
 Ambiguity: Ensure
meaningful to the user.
the
icons
are
For images, many users expressed an interest
in being able to ‘click on’ and enlarge the
pictorial image to full screen within
interactive television in order to clarify the
meaning of the image. Others suggested text
attributes for images similar to web-based
applications so that users have an alternative
to graphical information. Consider therefore:
 Alternative text: Appropriate textual
description of the content that would fit the
representation of the image or icon. Similar to
web based applications; alternative (ALT)
image tags could mark up the application
coding.
 Alternative image enlargement: The
possibility of allowing on-screen image
magnification functionality to compensate for
low vision
We advocate for having both methods of
accessing additional or enhanced content. The
user could then select which option was more
appropriate for them in a specific content
context. For example, if the user had a good
understanding of what the image represented,
such as in a news programme, alternative text
may not be necessary for that individual and
could be deselected.
Conclusion and further work
We have presented some of the limitations of
interactive television interface designs for
people with vision impairments. New design
strategies and models of interaction are
necessary to ensure that interactive television
is inclusive of this population of users. All
viewers including those with vision
impairments need to maintain a high level of
control of the digital broadcasting content
viewed, irrespective of physical restrictions,
and enjoy the experience on an equivalent
playing field as much as possible.
Further work on designing new interfaces that
can support skills, abilities and personal
preferences for a dynamic range of audience
groups is an important research direction.
This
includes
developing
adaptive
technologies to provide greater access to
interactive content, and understanding how
the content itself may actually shape the
interface design. Our challenge for visually
impaired audiences is to develop design
principles similar to those found in computing
and the human computer interaction field that
can support and accommodate a large range
of needs and user requirements.
References
1. Cunningham, C., and Coombs, N., 1997.
Information access and adaptive technology.
Phoenix, Arizona: The Oryx Press
2. Gill, J., 2000. Which button? Designing user
interfaces for people with visual impairments.
London: RNIB
3. Jacko, J, A., Dixon, A, M., Rosa, R, H.,
Scott., and Pappas ,C, J,. 1999. Visual profiles:
A Critical Component of Universal Access.
CHI ’99
Conference on Human factors in
Computing Systems (Pittsburgh, USA, May
1996) pp 330-337
4. Jackson, R., MacDonald, M., and Freeman,
K., 1994. Computer generated colour. A practical
guide to presentation and display. New York:
John Wilgey & Sons
5. Legge G, E., Rubin, G, S., and Luebker, A.,
1987. Psychophysics of reading. V. The role of
contrast in normal vision. Vision Research, 27,
pp 1165-1171
6. Murch, G, M., 1987 Colour graphics Blessing or Ballyhoo? Human-computer
interaction: a multidisciplinary approach.
Morgan Kaufmann Publishers Inc., San
Francisco, CA,
7. Peli E, Arend L, Labianca AT. 1996.
Contrast perception across changes in
luminance and spatial frequency. J Optical Soc
Am A 13(10): 1953-1959
8. Rice, M., 2003. Television and visual
impairment: Prospects for the accessibility of
interactive television. Proc: HCI International
2003, pp 800-804
9. RNIB., 2000. Review of the statutory
requirements for the provision of subtitling, sign
language and audio description services on digital
terrestrial television. London: Department
Culture, Media and Sport.
10. Strothotte, C., and Strothotte, T., 1991.
Seeing between the pixels. Pictures in interactive
systems. London: Springer
11. Vanderheiden, G., 1994. Application
software design guidelines: Increasing the
accessibility of application software to people
with
disabilities
and
older
users.
http://trace.wisc.edu/