A Method to Provide Accessibility for Visual Components to Vision Impaired

Azadeh Nazemi & Iain Murray
International Journal of Human Computer Interaction (IJHCI), Volume (4) : Issue (1) : 2013 54
A Method to Provide Accessibility for Visual Components to
Vision Impaired
Azadeh Nazemi azadeh.nazemi@postgrad.curtin.edu.au
Electrical and Computer Engineering Department
Curtin University
Perth, WA, Australia
Iain Murray i.murray@curtin.edu.au
Electrical and Computer Engineering Department
Curtin University
Perth, WA, Australia
Abstract
Non-textual graphical information (line graphs, bar charts, pie charts, etc.) are increasingly
pervasive in digital scientific literatures and business reports which enabling readers to easily
acquire the nature of the underlying information [1]. These graphical components are commonly
used to present data in an easy-to interpret way. Graphs are frequently used in economics,
mathematics and other scientific subjects. In general term data visualization techniques are
useless for blind people. Being unable to access graphical information easily is a major obstacle
to blind people in pursuing a scientific study and careers [2].This paper suggests a method to
extract implicit information of Bar chart, Pie chart, Line chart and math’s graph components of an
electronic document and present them to vision impaired users in audio format. The goal is to
provide simple to use, efficient, and available presentation schemes for non textual which can
help vision impaired users in comprehending form without needing any further devices or
equipments. A software application has been developed based on this research. The output of
application is a textual summary of the graphic including the core content of the hypothesized
intended message of the graphic designer. The textual summary of the graphic is then conveyed
to the user by Text to Speech software .The benefit of this approach is automatic providing the
user with the message and knowledge that one would gain from viewing the chart.
Keywords: K. Vision Impaired, Bar Chart, Pie Chart, Line Chart, Function Graph, Text to
Speech, GNUPLOT, GraphicReader.
.
1. INTRODUCTION
This Graphics are widely used in newspapers, text books, web pages, metro maps, instruction
manuals etc. They provide significant cognitive benefits over text. These graphical components
have an important role in conveying, clarifying, and simplifying information [3]. The majority of
information graphics that appear in formal reports, newspapers, and magazines are intended to
convey a message or communicative intention [4].Unfortunately, graphics are not easily accessed
by vision impaired people. In visual form, they are not accessible to vision impaired. Vision
impaired people unable to learn about the processes involved in reading, analysing, and
interpreting information presented in data visual graphs, tables and charts which are frequently
used in math and science materials to present and summarize data. It is fair to say that lack of
access to diagrams and other graphical content significantly limits educational and workplace
opportunities for people with vision impairment as well as access to popular media. This is in
contrast with textual content in which assistive technology have improved access [5]. Traditionally
charts are used to display trends and relationship, and also can be used to communicate
processes or display complicated data simply. These charts may be designed for the experts

trained users for data visualization or in popular media without complicated scientific reasoning
[6]. Charts are typically intended to convey a message that is an important part of the document
and this information generally not repeated in the article [7]. Thus people who are unable to see
and understand this chart will lose important part of information.
Several approaches have been addressed the accessibility of chart using alternative methods. A
number of projects have attempted to make graphic components accessible to vision impaired
users by reproducing the image in an alternative medium, such as sound [8], touch [9] or a
combination of the two [10][11].Viable alternatives include generation of a tactile graph, delivering
the information in text, interaction with an audio graph, or a combination tactile/audio approach
[12]. Non-textual components can be made accessible to the vision impaired in verbal description
or audio format and Haptic(tactual) feedback. These techniques help to express the data being
graphed in non visual ways. Synthesized speech and Braille are two common examples of
conveying information through the auditory and tactile senses [2]. Current approaches have
limitations such as the cost of translating into an accessible graphics format, use of expensive
tactile graphics or expensive peripheral devices, or lack of congruence with the original visual
graphic[5].Traditionally, graphs and diagrams are presented in Braille, or raised dots and lines on
the swell-paper [13]. Tactile graphics are images that use raised surfaces and vision impaired
user can feel them. They are used to convey non-textual information such as maps, paintings,
graphs and diagrams. Tactile graphics can be seen as a subset of accessible images [14].
Several problems are associated with Braille and Tactile technique:
• Only a small proportion of blind people can use Braille, because reading it requires
sufficient tactile sensitivity which all vision impaired do not have [15].
• Blind people can only get a rough idea about the content [13].
• Tactile diagrams are not durable. It is not easy to make changes to tactile diagrams.
• Reheating the swell paper could overcook the already raised parts and reduce the
quality [16].
• In the method ,making graphs tactile Picture in Flash (PIAF)[21], the general shape of the
graph can be understood by touching it carefully, but hardware is needed to generate
tactile charts. Two main effective restriction factors in this method which must be
considered for tactile symbols in charts are :Discriminating ability and Searching ability
[22]. Tactile symbols without these properties could not help to explore concepts of
charts by vision impaired people.
In addition, Haptic feedback which is more useful for guidance and assisting users’ navigation on
the graphs is not efficient to present exact data values to the user. Moreover, if too much
information is presented via Haptics, the narrow bandwidth can be easily overloaded [2]. Also it
may take users some time to familiarize themselves with the new interface. The limitations of the
force feedback devices also hinder users exploration on the graphs[18]. Vocalizing the form of the
graph and individual data is useful as long as a screen reader is provided to interpret trends in
the data, however, the shape of the chart has to be imagined in the user’s mind[20].
In addition other methods have some limitation such as:
• Conveying data via sound pitch and 2-dimensional acoustics[19].It is difficult to convey
data accurately with the acoustic method(non speech sound), and moreover, since
acoustics are volatile, information can easily be misheard.

• The Interactive SIGHT (Summarizing Information GrapHics Textually) system provides
visually impaired individuals with the high-level knowledge that one would gain from
viewing. SIGHT uses image processing techniques to extract communicative signals from
a chart, but it is still limited to present information of the bar chart within the WebPages
[4].
2. METHODOLOGY
There are two categories for methods to present visual component to vision impaired:
• Passive (description tool):The user is presented with a representation of the entire visual
part at one time, with limited user input.
• Active (exploration tool):The user can explore non textual part [23].
The offered method in this paper has been attempted to combine two mentioned categories. It
means this method presents a completed description of illustration (passive) and provides
navigation ability through various field of data to give user opportunity to explore and build a
mental map of the visual components (active).
On the other hand, providing an alternative equivalent for inaccessible content is one of the
primary ways authors can make their documents accessible to people with disabilities [18],
providing a hierarchy of text equivalents can also convey the hierarchical structure of the graphic
components [3]. The alternative content fulfills essentially the same function or purpose for users
with disabilities as the primary content does for users without any disability. Text equivalents are
always required for graphic information [3]. Therefore this research has focused to extract these
text equivalents from graphic components where they are not provided by authors.
For implementation this method, accurate understanding of the ways in which non-textual parts
benefit sighted people, together with the problems encountered when exploring them non-
visually, are required [24]. Also understanding of nature, meaning and concepts and obtaining
comprehensive information from graphics must be considered to present detailed descriptions.
The factors which appear to affect graphical comprehension are the purposes for using graphs,
task characteristics, discipline characteristics, the differences between presenting information
visually and aurally, and how and what graphical parts are represented in the mind[25]?
Most significant benefits for readers of visual parts are reducing demands on short-term memory
and gaining a quick overview [26] which have been considered in aural representation in this
method.
In general term each chart structure contains two parts:
1. Text part including :the number and amount of each axis ticks, axis labels ,legends and
title.
2. Image part including: graph body.[22].
To provide non textual graphical components accessibility to vision impaired users it is essential
to obtain comprehensive information from image .In addition advanced techniques for the graphic
understanding are needed [27].Since large quantities of data are presented in a graphic, thus
efficient access strategies are required according to the tasks that users are trying to accomplish
during navigation through information.
2.1 GraphicReader Software Application
GraphicReader is a software application which has been developed in order to provide real-time
and independent access to the vast repository of information in the graphics (bar chart, line chart,

pie chart and math’s graph), with automated image processing, math processing, decoding, and
extracting information which these graphical components convey and emphasis. After extracting
data as text, the system analyze them to generate XML format which is convertible to audio
format and used to provide easy navigation ability to vision impaired people .The application
integrates appropriate representation for information accessibility.
This main core is Visual Extraction Module (VEM) and regarding the charts structure contains two
Sub-Modules:
1. Text data extraction by performing accurate Optical Character Recognition (OCR) .Before
OCR some preparation steps must be done to clean the text background and enhance
the text .Theses processing steps are:
• 90 degree rotate if aspect does not match layout
• Crop
• convert to grayscale
• enhance filter to clean background and remove noise
• text smoothing
• Sharpening
• saturation change
• alternate text smoothing via adaptive blur
After OCR , running Text Analyzing Module or Data Interpretation leads to gain access to chart
title, axes labels, the number and value of ticks ,and number of slices.
2. Image data extraction contains graph digitization to provide data table .Data table is a
table with two columns to shows amount of each dependent variable related to its
independent variable.
After data extraction and data collection, XML generating Module generates XML code based on
all collected data and considering the result of Data Interpretation Module. The availability of
markup format for the description of a chart helps to concentrate on the development of
presentation strategies which comprehensively expresses the original chart in audio format
without needing further peripherals.
By completing XML code user is passively presented key statistical information about the
presented chart. For users who do not need detailed description, an static description should be
generated for a quick access.
Since any technique that aims to provide non-visual accessibility of charts needs to support the
processes of mental model formation and manipulation, at this stage user can easily navigate the
chart and has the benefit of both passive and active presentation .It means the user is guided by
the system in navigating the content of the chart or can control the presentation by deciding what
parts of the charts to focus on. The main important features of navigation ability in
GraphicReader, are reducing the demands on memory and reassuring user to explore the chart
meaning.

International Journal of Human Computer Interaction (IJHCI), Volume
Thus depending on user navigation request which is s
different classified data could be extracted from XML representation, converted to plain text and
sent to TTS module to make audio representation.
In addition XML representation is converted to a .dat file and .gnu co
GNUPLOT module to generate accessible graphical component using classified data in XML
representation. In fact this module is a feedback module and by comparing graphical component
produced by GNUPLOT and original image, accura
found and corrected. Figure 1 illustrates the overall architecture of the GraphicReader access
system.
FIGURE 1
2.2 GNUPLOT
GNUPLOT is an open source command
dimensional plots of functions, data, and data fits. It is frequently used for publication
graphics as well as education. One major accessibility benefit derived from GNUPLOT is
encoded as plain text. Authors can cr
text encoding also means that people may use relatively simple, text
with visual impairments, and can be used to supplement graphical rendering. GNUPLOT presents
prepared, simple, text based and easy to use pattern to generate math’s function graph, bar char
and pie chart.
3. EXAMINED VISUAL COMPONENT
3.1 Bar Chart
The Bar charts are one of the most commonly used visualization techniques and are often
encountered daily, e.g. on newspapers, journals and magazines. Bar chart contains rectangles
International Journal of Human Computer Interaction (IJHCI), Volume (4) : Issue (1) : 2013
Thus depending on user navigation request which is send to system through user interface
sent to TTS module to make audio representation.
In addition XML representation is converted to a .dat file and .gnu configuration file and sent to
produced by GNUPLOT and original image, accuracy of system can be estimated and errors are
FIGURE 1: GraphicReader access system flowchart.
GNUPLOT is an open source command-line program that can generate two
data, and data fits. It is frequently used for publication
encoded as plain text. Authors can create and edit it with a text-processor authoring tool. Plain
text encoding also means that people may use relatively simple, text-based. This can help users
ed, simple, text based and easy to use pattern to generate math’s function graph, bar char
EXAMINED VISUAL COMPONENTS
g. on newspapers, journals and magazines. Bar chart contains rectangles
58
end to system through user interface
nfiguration file and sent to
cy of system can be estimated and errors are
that can generate two- and three-
data, and data fits. It is frequently used for publication-quality
processor authoring tool. Plain
based. This can help users
ed, simple, text based and easy to use pattern to generate math’s function graph, bar char
g. on newspapers, journals and magazines. Bar chart contains rectangles

which the height indicates the quantities of the corresponding variables. They usually show
discrete and independent variables.
For a bar chart, the representation includes the numbe
axis, and information for each bar such as its label, the height of the bar. Some emphasized
information implicit in bar chart are:
• The highest bar on the graph.
• The lowest bar on the graph
• Two bars which have the closest values.
GraphicReader currently handles electronic images contain Bar chart produced with no
overlapping characters. Also it hypothesizes standard placement of labels and axis headings.
VEM for Bar chart contains several image processing
In order to extract vertical and horizontal axis labels and number of bars image must be sent to
OCR engine.
Figure 2 illustrates the sample of a bar chart which indicates the “Average Price of
Advertisements of five different chan
,extracted text from OCR.
FIGURE 2: Sample of Bar chart, mono color image, vertical lines image
Output Text file generated by OCR helps to recognize number of bars (in this example is five)
labels of bars (NBC, ABC, FOX, CBS, WB
by step value 40). By converting the chart to grey scale mono graph, removing horizont
splitting image of vertical lines to tiles and using histogram of each tile, number of black pixels in
each bar has been obtained, the height of each bar can be calculated by comparison with vertical
axis maximum value and its number of pixels. F
black pixels, thus the value of NBC bar by having
between value of each bar and its number of black pixels.
discrete and independent variables.
For a bar chart, the representation includes the number of bars in the graph, the labels of the
information implicit in bar chart are:
The highest bar on the graph.
The lowest bar on the graph
the closest values.[16].
VEM for Bar chart contains several image processing steps such as OCR.
nts of five different channels Grey scale mono graph ,image of vertical lines
Sample of Bar chart, mono color image, vertical lines image, , OCR output
generated by OCR helps to recognize number of bars (in this example is five)
, ABC, FOX, CBS, WB) and scaling value on vertical axis(between 0 to 200
by step value 40). By converting the chart to grey scale mono graph, removing horizont
axis maximum value and its number of pixels. For example vertical axis has value 200 by 177
black pixels, thus the value of NBC bar by having 153 black pixels is 172 .Table1
between value of each bar and its number of black pixels.
200
160
120
80
40
0
NBC ABC FOX CBS WB
59
r of bars in the graph, the labels of the
image of vertical lines and
, OCR output.
generated by OCR helps to recognize number of bars (in this example is five),
) and scaling value on vertical axis(between 0 to 200
by step value 40). By converting the chart to grey scale mono graph, removing horizontal lines,
or example vertical axis has value 200 by 177
shows relation
NBC ABC FOX CBS WB

Tile no No of black pixels Height of Bar
0 330 Vertical axis +NBC
1 153 NBC=200*153/177=172
2 147 ABC=2000*147/177=166
3 147
4 106 FOX=200*106/177=120
5 106
6 76 CBS=200*76/177=86
7 76
8 34 WB=200*34/17=39
9 34
TABLE 1: Height of bars in the sample bar chart
The following is XML codes generated by XML generating Module using data extraction from
VEM:
<bar-chart>
<horizontal-axis-label>”Channel” </horizontal-axis-label>
<vertical-axis-label>”Millions of Dollars” </vertical-axis-label>
< vertical-axis-range >”0:200” < /vertical -axis-range >
<vertical-axis-tic>”40” </vertical-axis-tic>
<x,y>NBC,172</x,y>
<x,y>ABC,166</x,y>
<x,y>FOX,120</x,y>
<x,y>CBS,86</x,y>
<x,y><WB,39</x,y>
</bar-chart>
After XML generation, the system produces .gnu configuration file and .dat file to provide
GNUPLOT package required files such as the following:
#Bar.gnu
reset
set output "bar.png"
set ylabel " Dollars"
set ytics "40"
set style fill solid
set term png truecolor
set xlabel "Channel"
set yrange [0:200 ]
set boxwidth 0.5
plot "data.dat" using 1:3:xtic(2) with boxes
#Bar.dat
0 NBC 172
1 ABC 166
2 FOX 120
3 CBS 86
4 WB 39
Figure 3 shows an accessible image which is output result of GNUPLOT package. This image
can be replaced by original image for more usage and getting more efficient and quicker results
in terms of accessibility. Comparison this output image and inaccessible original input image has
shown a little difference in value of bars height.

FIGURE 3
3.2 Pie Chart
A pie chart is a circular chart divided into
length of each sector (and consequently its
it represents [28] The symbols are the pie sectors. Each pie sector has to be matched with the
variable via the legends. By the matching algorithm, the underlying data can be interpreted. VEM
for pie chart supposes which legend information are in descending order.
Figure 4 illustrates the sample of a pie chart which indicates “Annual Market Share Analysis for
three different products“, extracted text from OCR and color histogram information extracted by
image processing.
FIGURE4: Sample of Pie chart, OCR output,
As it is observed from histogram information the total number of colored pixels is
17029+12654+10898=40581.Table 2 has shown the percentage of each product.
Slot
Product A
Product B
Product C
TABLE 2
XML codes generated by Marks up Format Generating Module are as the following:
<title>” Annual Market Share A
< slot
FIGURE 3:GNUPLOT output Bar chart.
divided into sectors, illustrating proportion. In a pie chart, the
of each sector (and consequently its central angle and area), is proportional
The symbols are the pie sectors. Each pie sector has to be matched with the
t supposes which legend information are in descending order.
Product A
Product B
Product C
17029: (255, 52, 1,255) #FF3401
rgba(255,52,1,1)
12654: ( 53,101,201,255) #3565C9
rgba(53,101,201,1)
10898: (255,154, 2,255) #FF9A02
rgba(255,154,2,1)
Sample of Pie chart, OCR output, Histogram output.
As it is observed from histogram information the total number of colored pixels is
0898=40581.Table 2 has shown the percentage of each product.
Percentage Accuracy level
17029*100/40581=41.96% 99.4%
12654*100/40581=31%6. 98.7%
10898*100/40582=26.85% 98.7%
TABLE 2: Slots percentage of sample pie chart.
<pie-chart>
<title>” Annual Market Share Analysis for three different products“” </title>
<no of slots l>”3” </no of slots>
< slot-1 >”Product A,41.96%” < /slot-1 >
61
sectors, illustrating proportion. In a pie chart, the arc
to the quantity
The symbols are the pie sectors. Each pie sector has to be matched with the
17029: (255, 52, 1,255) #FF3401
12654: ( 53,101,201,255) #3565C9
10898: (255,154, 2,255) #FF9A02
Accuracy level
99.4%
98.7%
98.7%
nalysis for three different products“” </title>

< slot-2 >”Product B,31%” < /slot-2 >
< slot-3 >”Product C,26.85%” < /slot-3 >
</pie-chart>
3.3 Line Chart
Line chart displays information as a series of data points connected by straight line segments
[29]. It is a basic type of chart common in many fields. It is created by connecting a series of
points that represent individual measurements with line segments. A line chart is often used to
visualize a trend in data over intervals of time [30]. A line chart is typically drawn by two crossed
lines, called axes. The horizontal axis is called the x-axis which represents the independent
variable and the vertical axis is called the y-axis which represents the dependent variable. The
line chart may contain additional lines drawn parallel to both axes which is called grid. Each axis
indicates significant values with small marks, called ticks. A short description of the axis is label
The chart overall description is title.
For line chart, Graph digitization is a process involves converting the pixels in the image of line
chart to original (x, y) data values. Extracted original (x, y) data values from the image helps to
generate digitizing data table [31].
Therefore graph digitization has two responsibility, First finding the position of pixels of the main
graph in the image and then running mathematical function for all found pixels to assign a
coordinate system to the pixels in the graph based upon the extracted scaling values, line chart
image size ,and line chart origin point pixel position. Although the basic concept of converting the
image pixel values to scaled values is straightforward, practical consideration such as specifying
accurate origin point must be addressed.
• Finding graph pixels position process includes several image processing steps as the
following :
Resizing image to 640x480
Removing all gridlines it is essential to discriminate grid lines from the axes.
Obtaining graph image without axes and text by using color separation and color
histogram information
Conversion graph only image to grayscale
Flipping image to move origin point from top left corner to bottom left corner
Negation graph image
Converting image to a text file including information of each pixel location
Removing all black pixels information from text file .Remaining pixels are body
graph pixels
Sorting data in text file based on first column(x values)
• Conversion graph pixels position to original (x, y) data value depends on finding origin
point pixel position, obtaining X-ratio and Y-ratio. To approach this goal, following step must
be done:
Obtaining axes image without graph using color separation and color histogram
information
Removing axes ticks and axes labels and generating an image which contains
only horizontal axe.
Converting horizontal axe image to a text file including information of each pixel
location in horizontal line
Removing all white pixels information from text file related to horizontal axe
Obtaining position of origin by finding Minimum X value and Maximum Y
value.(which Xo, Yo are origin coordinate values, and W,H are weight and
height pixel sizes of image )
Xo= Minimum_X_ value
Yo=480- Maximum_ Y_ value
W= Maximum_X_ value- Minimum_X_ value+1

H= Maximum_Y_ value- Minimum_ value+1
X-ratio=amount of X_ticks*number of X_ticks/W= Max-x-Scale /W
Y-ratio=amount of Y_ticks*number of Y_ticks/H= Max-y-Scale /H
Conversion pixel to data has been done using following formulae:
(which Xp,Yp are pixel position coordinate values, and X,Y are original value in
chart)
X= (Xp –X`o)*X_ratio
Y= (Yp-Yo)*Y_ratio
Converting X floating values to integer using function “floor(x)” as the following :
If x<0: then return int(x)-1
If x>0: then return int(x)
Function floor(x) is used except for last X value which is used “ceil(x)” function to
convert to integer.
If x<0: then return int(x)
If x>0: then return int(x) +1
Converting Y floating values to integer using function round(y)
After performing all these processes , digitizing data table of different value for X ,Y has been
generated and with presenting numerical data in the form of a line chart relative values ,relative
ratios and interpretation of trends , can be understood instantly. Also it should be useful in
addition X, Y, slope of the individual line between two adjacent points has been calculated and
added to data table. It should be calculate from this formula:
Slope_i= ((ܻ௜-ܻ୧ିଵ)/ ሺܺ௜ିܺ୧ିଵ))
Calculating slope between two points helps to provide better summary description for line chart
trends using the following rules:
If a line segment slopes upward from left to right, the slope is positive.
If a line segment slopes downward from left to right, the slope is negative.
If a line segment is horizontal, the slope is 0 (zero).
If a line segment is vertical, the slope is undefined.
Most considerable factors which must be described in a line chart are as the following:
• Counting the number of lines.
• Identifying the line friction.
• Counting the number of bends on the line
• Counting the number of intersections
Table 3, indicates a table of original values which generate the sample line chart.
TABLE 3: Original values of the sample line chart.
Figure 5 illustrates the sample line chart generated by above
X Y
0 5
1 7
2 3
3 11
4 15
5 13
6 9
7 14

FIGURE 5: The sample line chart .
The results of first VEM sub module are :
Data extracted Module 1:Max-x-Scale=10 ,Max-y-Scale=15
The result of second VEM sub module contain origin point pixel position and graph dimension
are: Xo=60 Yo=22 W=547 H=420
And finally table 4 shows the result of Graph Digitization before converting to integer values:
X Y
0.18 5
1.05 6.92
2.10 3.035
3.16 10.96
4.03 14.85
5.09 12.89
6.14 9
6.98 13.98
TABLE 4: Data extracted from line chart by GraphicReader.
The generated XML codes regarding extracted data table from sample line chart comes below:
<line-chart>
<title>” “</title>
<horizontal-axe-label>”X” </horizontal-axe-label>
<vertical-axe-label>”Y” </vertical-axe-label>
< horizontal-axe-range >” 0: 10” < /horizontal-axe-range >
< vertical-axe-range >” 0 : 15” < /vertical -axe-range >
< x_tick>”1”</x_tick>
< y_tick>”1”</y_tick>
<x0,y0>”0,5”</x0,y0>
<x1,y1><”1,7”></x1,y1>
……
<x7,y7><”7,14”></x7,y7>
<slope0>”2”<.slope0>
<slope1>”-4”<.slope1>
….
</line-chart>
The application provides users the opportunity to select several presentation options such as
chart title, average point, maximum point, minimum point, start point, endpoint, or trend of the
chart points. The user provides input to the navigation process through keyboard controls.
3.4 Graph of math’s function
Graphs are visual presentation for math information and useful to illustrate qualitative information,
which are sometimes even impossible to describe them .A graph is a picture that shows how sets

of data are related to each other. Generally structure framework of graph of the function contains
two crossed axes which meet on origin point. The axes divide the plane into four quadrants.
Scaled values are increased from left to right on horizontal axe and respectively from down to up
on vertical axe. For describing graph of the function labels , scaled value and the range of each
axes, where the function grows or decreases, where the maximums, the minimums and the
flexes are and where the points of discontinuity are must be considered. Also following factors
should be included in graph description [14][16].
• Meet points of the graph with axes: These points are obtained by assigning zero to “x “and
finding “y “value and assigning zero to” y” and finding “x”. This step requires solving the
equation .For example one method for this purpose is using :
Quadratic Formula : ax2 + bx + c = 0 value of x is given by:
‫ݔ‬ ൌ
െܾ േ √ܾଶ െ 4ܽܿ
2ܽ
• Inflection point: The function is continuous at the point and the concavity of the graph
changes at that point. This point is found by make the y”, zero.
• Starting and ending points: These points are obtained by assigning minimum and
maximum values indicated on vertical axe to y and finding x.
• Minimum and maximum points: These points are obtained by assigning zero to y’ and
finding x at this point :
If y “>0, then the point is maximum
If y “<0, then the point is minimum
If y”=0 ,then the point can be both or neither
• Shape: Regardless details, shape of graph function depends on function, so in most cases
general shape of graph can be guessed by having function. Some rules to find shape of
graph are as the following:
If y=ax+b ,then the shape is a line
If y=ax2 + bx + c or quadratic functions ,then shape of graph is parabola
If 'a' is positive, the graph will be 'U' shaped. If 'a' is negative, the graph will be 'n'
shaped. The graph will always cross the y-axis at the point c (so c is the y
intercept point).
If for an specific interval x2>x1 and y2>y1 ,or y’>0 ,then in this interval graph is
increasing
If for an specific interval x2>x1 and y2<y1 ,or y’<0 ,then in this interval graph is
decreasing
If for an specific interval y’=0,then in this interval graph is constant
If f(x)=f(-x) ,then graph has symmetry regarding y or even
If f(x)=-f(-x) ,then graph has symmetry regarding origin or odd
If y”>0 then graph is concave up
If y”<0 then graph is concave down

Figure 6 illustrates a sample for graph function using y=x^2+2*x as function formula.
FIGURE 6. Sample of math’s graph function sample.
Table 5 is the data table of above graph obtained from performing Data Digitization using
mathematic processing.
TABLE : Extracted data from math’s graph
function.
XML codes are generated by GraphicReader using information extracted from mathematical
analyzing.:
</graph>
<shape>”parabola”</shape>
<horizontal-axis-label>”X” </horizontal-axis-label>
<vertical-axis-label>”Y” </vertical-axis-label>
< horizontal-axis-range >”[-3:3]” < /horizontal-axis-range >
< vertical-axis-range >”[-3:3]” < /vertical -axis-range >
<horizontal-axis-tic>”1” </horizontal-axis-tic>
<vertical-axis-tic>”1” </vertical-axis-tic>
<equation>”x**2-2*x”</equation>
<vertical -intercept>”0,0”</vertical-intercept>
<horizontal-intercept>”0,0”</horizontal-intercept>
<horizontal-intercept>”2,0”</horizontal-intercept>
<starting -point>”-1,3”</starting -point>
<ending-point >”3,3”</ending-point>
/<minimum-point>”1,-1”</minimum-point>
</graph>
Description results of GraphicReader application using mathematical analyzing and data s
• Horizontal axis is labeled X and its range is between -3 to 3
• Vertical axis is labeled Y and its range is between -3 to 3
• Horizontal and vertical step value is 1
• This graph is plotted from the equation: y= x^2 -2x
• The graph is started from point (-1, 3) from top left quarter of plane until (3,3)
located on top right quarter of plane :
x^2 -2x=3 x=3 x=-1
X Y
-3 15
-2 8
-1 3
0 0
1 -1
2 0
3 3

x^2-2x=-3 no answer
• Horizontal intercept occurs at points 0,0 and 2,0(y=0)
• Vertical intercept occurs at point 0, 0 (x=0)
• Shape of graph is parabola and has minimum (a=1>0) value at (if y’=0 2x-2=0,
then x=1, y=-1) point (1,-1 )
4. COMAPRITIVE EVALUATION
Comparison between this research method (using synthesis speech) and tactile representation
for non textual components to vision impaired users shows that tactile representation requires
manual preparation, a peripheral device and user training . Additionally, these original visual
components must have some properties that may be transposed to a tactile representation thus
limiting the possible attributes that may be included. For example;
• Pie charts must be enlarged to display the divisions of the circle. If numbers are shown,
they should be preceded by the numeric indicator (number sign) regardless of whether
the text is transcribed in literary or Nemeth (mathematics and science) code. The outline
of the circle is tactually difficult to discern from the lines separating the divisions. A few of
these lines may be stopped just short of the centre to avoid clutter. [34].
• Tactile representations for bar charts provide relative comparison of magnitude, not
detailed, accurate values.
• Using tactile methods to access information in line charts provides presentation without
measurement or scale. It may be difficult to differentiate between division lines that are
close together.
• Tactile method for Math’s graph only represents the shape of graph without any detail.
Non-speech methods, when compared with synthesis speech, provides an overview of
the data, as it can be delivered in a shorter time than synthetic speech. Encoding the
data value in the pitch of a MIDI note is a common strategy used to represent relative
values of numerical data. A user can scan the pie chart or math’s graph to get an idea of
the overall distribution of the data, without accessing to the exact values [33].
• Synthetic speech which is used in this research, delivers detailed information such as
actual values and proportion to vision impaired users. It can be used to represent text
labels and precise data values without cluttering the tactile image with Braille or raised
letter components. However listening and comparing many values passively through
speech alone can be laborious and time consuming, considering navigation through
information and active representation solves the issue regarding sequential passive
representation.
• Previous researches show accuracy for non speech audio method to transfer information
of line chart is approximately 80%[12] and for virtual touch (Haptic) is between 85.56%
and 94.07% [32]. This method utilized in this research showed an accuracy in conveying
implicit data in line chart at approximately 98% and for pie charts, 98.7%.
5. CONCLUSION AND FUTURE WORKS
This research, has designed a system to provide non-visual aural access to charts. The proposed
approach has advantage of the ability to extract substantial information of charts from the image.
Since the Visual Extraction Module assumes that it is dealing with a simple chart created by
“GNUPLOT” and title position, axes ticks and, axes label locations must be in “GNUPLOT” forms,
further work is required to extend this work to other charts layouts and digital formats. Issues
remain with the user interaction and making the most appropriate and efficient description for
effective communication of graphical information in documents to vision impaired

6. REFERENCES
1. J. Gao , Y. Zhou, K.E. Barner,” Making Information Graphics Accessible for Visually Impaired
Individuals” ,University of Delaware Newark ,DE19716 USA, March 2012
2. W. Yui, S. Brewster, “Evaluation of Multimodal Graphs for Blind People” , Glasgow,2003
3. C. McCathie Nevile, M.R.Koivunen, Accessibility Features of SVG, Available at :
www.w3.org/TR/SVG-access, Aug 2000,[ Nov.1, 2012].
4. S Elzer, E Schwartz, S Carberry, D Chester, Demir, P Wu,“A browser extension for providing
visually impaired users access to the content of bar charts on the web”, Millersville University,
POB 1002, Millersville, PA 17551, USA,2007
5. Accessible Graphics: Graphics for Vision Impaired People,Diagram 2012 7th International
Conference on the theory and application ,2012
6. C. F. Greenbacker, P.Wu S. Carberry, K.E. McCoy,S. Elzer ,D. D. McDonald ,D. Chester, S.
Demir. ”Improving the Accessibility of Line Graphs in Multimodal Documents” Published in:
Proceeding SLPAT '11 Proceedings of the Second Workshop on Speech and Language
Processing for Assistive Technologies,2011
7. S Carberry, S. Elzer, S. Demir.” Information graphics: an untapped resource for digital libraries”
Proceedings of the 29th Annual International ACM SIGIR Conference on Research, 2006
8. P.B.Meijer, An experimental system for auditory image representations, IEEE Transactions, on
Biomedical Engineering, 39(2):112–121. 1992
9. S. Ina,”Computer graphics for the blind” ACM SIGCAPH Computers and the Physically
Handicapped,55:16–23,1996
10. R.Ramloll,W. Yu, S. Brewster,B. Riedel, M. Murton,G. Dimigen, “ Constructing
sonified haptic line graphs for the blind student:First steps”, Proc. of ASSETS 2000, p. 17–
25,Arlington, VA, 2000
11. P.Roth, D Richoz, L. Petrucci, T. Pun,”An audio-haptic tool for non-visual image
representation” ,Computer Science Department CUI, University of Geneva, XCH – 1211 Geneva
4, Switzerland,2001
12. National Centre for Accessible Media (NCAM),Provide access to graphs for users who are
blind or visually impaired, Available at :
http://ncam.wgbh.org/invent_build/web_multimedia/accessible-digital-media-guide/guideline-f-
graphs, [ 2 Dec 2012]
13. W. Yu, S. Brewster, “Multimodal Virtual Reality Versus Printed Medium in Visualization for
Blind People”, Glasgow,2002
14. D.McCallum,”The design and manufacture of tactile maps using an ink-jet process”, Journal
of Engineering Design, ,2005
15. I.Murray,” Instructional eLearning technologies for the vision impaired “.published doctoral
thesis, 2008. Curtin University, Perth,2008
16. W.Yu,R. Ramloll,S.Brewster,” Haptic Graphs for Blind Computer Users”, Glasgow,2001
17. A. R. Kenne”,Audiograf: A diagram-reader for the blind”. In Second Annual ACM Conference

on Assistive Technologies, p. 51–56.1996
18. University of South Australia ,2004,”Online Accessibility Action Plan” ,Available at:
http://w3.unisa.edu.au/footer/accessibility/actionplan.pdf,[ 20 Nov 2012]
19. L. Brown, S. Brewster, R. Ramboll, M. Burton and B. Riedel ,”Design guidelines for audio
presentation of graphs and tables” ,Proc. of ICAD 2003, 2003.
20. L. Ferres, P. Verkhogliad, G. Lindgaard, L. Boucher, A. Chretien, and M. Lachance,
“Improving accessibility to statistical graphs: The iGraph-Lite system”, ASSETS’07, Proc. of 9th
Int. ACM SIGACCESS Conf,2007
21. C. Goncu , K. Marriott, “Tactile chart generation tool”, ASSETS’08, Proc. of 10th Int. ACM
SIGACCESS Conf. Computers and Accessibility, pp.255-256, Halifax, Nova Scotia, Canada,
October 2008.
22. T.Watanabe, T. Yamaguchi, M. Nakagawa, “Development of Tactile Graph Automated
Creation Software”, University of Niigata, Japan,2012
23. R Cohen, V. Haven, J.A. Lanzoni, A..Meacham,”Using an Audio Interface to Assist Users who
are Visually Impaired with Steering Tasks”,Boston, 2006
24. A. Brown, R. Stevens ,S Pettifer, “Making graph-based diagrams work in sound: the role of
annotation”, Manchester,2012
25. S. Friel, F.R. Curcio, G.W. Bright,” Making Sense of Graphs:Critical Factors Influencing
Comprehension and Instructional Implications”, Journal for Research in Mathematics Education
2001, Vol. 32, No. 2, 124–158
26. A.J. Brown,” Non-Visual interaction with graphs”, Manchester, 2007
27. W. Huang,C.L. Tan,” A system for understanding imaged info graphics and its applications” ,
Proceedings of the 2007 ACM symposium on Document engineering,2007
28. W.S.Cleveland. The Elements of Graphing Data. Pacific Grove, CA: Wadsworth & Advanced
Book Program,1985
29. Burton G. Andreas . Experimental psychology,1965
30. N J. Salkind Statistics for People who (think They) Hate Statistics, The Excel Edition,2006
31. D.L.Ryan,” Modern Graphic Communications” ,Prentice-Hall, Englewood Cliffs, NJ, 1986
32. W,Yu, R Ramloll, S Brewster, B Ridel,”Exploring computer-generated line graphs through
virtual touch”Department of Computing Science, *Department of Psychology University of
Glasgow, G12 8QQ, UK,2012
33. S. A. Wall , S. A. Brewster “Tac-tiles: Multimodal Pie Charts for Visually Impaired Users”
Glasgow Interactive Systems Group, Department of Computing Science University of Glasgow,
G12 8QQ, UK,2012
34. Guidelines and Standards for tactile Graphics .available at http://brailleauthority.org/tg/web-
manual/, 2010,[3 Dec 2012]

A Method to Provide Accessibility for Visual Components to Vision Impaired

More Related Content

Viewers also liked (18)

Similar to A Method to Provide Accessibility for Visual Components to Vision Impaired (20)

More from Waqas Tariq (20)

Recently uploaded (20)

A Method to Provide Accessibility for Visual Components to Vision Impaired