Norbert Kuhn, Stefan Richter and Stefan Naumann
Institute for Software Systems in Business, Environment, and Administration
University of Applied Sciences Trier, Location Birkenfeld, P.O. Box 1380, D-55761 Birkenfeld, Germany
Keywords: Workflow Management, Printed Forms Processing, Accessibility, Assistive Technology.
Abstract: Although many companies and governmental institutions have provided their costumers with access to
electronic information systems and to their processes, there still remains widespread use of paper-based
communication. In this paper we present an approach, which we use in the FABEGG system to make these
documents accessible to handicapped users. We discuss a prototypical system with both an administrative
and a user front-end to work with printed forms in business processes. Compared to purely electronic
solutions we can also cope with forms, which contain already process related data like for instance name,
address and other personal information of the recipient. The prototypes presented here will be evaluated in
local governmental authorities.
In recent years much effort has been spent in Human
Computer Interfaces to improve access for
handicapped persons to computer systems (Mulleret
al. 1997). To a major extent these activities are
enforced by legislative constraints that exist in the
US (e.g. the Americans with Disabilities Act (United
States of America, 1990)) as well as in the European
Union (European Commission, 2000), and in its
member countries, like in Germany (Bundesrepublik
Deutschland, 2006).
In this paper we focus on the interfaces to
information systems for humans with visual
impairments. In general, there exists a variety of
hardware or software components to enable access
to computer systems for this group of users (for
early approaches see e.g. (Edwards, A. D. 1988,
Ballenger, W. L. 1979)). These tools are either
delivered as part of the operation system, e.g. screen
magnifiers in the Windows operation systems, or by
companies specialized in assistive technologies, e.g.
ScreenReader, Braille displays and text-to-speech
software, or they may be included as an additional
feature of a standard software system, like the
speech input/output facility in the Opera Internet
browser. We refer to http://www.barrierefrei- for an overview of such tools.
For the design and implementation of such software
systems there exist guidelines, frameworks and
tools, e.g. (Sun Microsystems, 2006), (W3C, 2006).
These tools improve accessibility on the client side
for visually impaired users. However, they do not
consider the complete activity of information
processing. For instance, accessibility can be
improved by including additional information within
the document of interest or by special classification
algorithms to classify these forms.
From a process related point of view we may
observe that particularly governmental institutions
have been active to be present also in the Internet
and they have started to provide their services for
electronic access. However, when looking closer at
this development we may recognize that many of
these realizations only allow the citizens to
download particular forms, usually in PDF-format,
that have to be printed, filled and sent back to the
governmental institution. Furthermore, when an
institution contacts a citizen this usually happens by
paper mail. Often, these documents are forms that
are partially filled with personal data of the recipient
who is in charge to complete the form. It is obvious
that this process step is difficult to handle for
visually impaired people, elderly or dyslexic people,
or immigrants.
In the following we present an approach to
improve this situation by exploiting different
Kuhn N., Richter S. and Naumann S. (2007).
In Proceedings of the Ninth International Conference on Enterprise Information Systems - HCI, pages 286-289
DOI: 10.5220/0002374802860289
technologies e.g. from the fields of document
analysis and language processing.
The FABEGG system consists of different
components. The most important ones for the
underlying paper are on the one hand the component
that allows construction of a document repository by
a company or a governmental authority. We will
stretch the possibility to insert forms into the
repository and hence, refer to this component as the
forms front-end hereafter. The second component is
the one that enables the use of the repository and
will be referred to as user front-end in the following.
The document insertion is organized as follows:
A governmental authority feeds documents and
forms into a document and template repository via
the FABEGG forms front-end. For the discussion
within this paper we mainly focus on the insertion of
forms that are part of processes within the authority.
The FABEGG forms front-end allows the authority
to enhance the forms with metadata that is necessary
to process them later on. Furthermore, it allows
adding metadata to support visual impaired users to
understand and to process these forms.
The FABEGG user front-end builds on this
information to present documents. The user can scan
a document to process and he will then have
different modes of access to it. For example the
document can be displayed in certain colour
combinations, in an enlarged mode or it can be read
to the user (cf. section 5).
In the next sections we will describe these two
components in more detail.
FABEGG provides an interface to insert forms or
general documents into a repository. This repository
can be used like an information centre for a
company or a governmental institution where
relevant information can be found and downloaded.
In the following we focus on the insertion of forms
that are usually presented to the citizens as printed
documents. For digital forms a similar procedure is
In order to guide a blind or visual impaired user
through the form a set of help texts can be associated
with the document. This comprises text to explain
the general processing steps of the form as well as
help texts for each field to explain the user the kind
of information that is expected to complete a field.
To enable the forms recognition step from the
digital image a set of anchors is associated with a
document or form. In general, anchors are areas
where certain content is expected or which have
special features that allow identifying the form or
the form template, respectively in the repository.
Documents and templates are contained in the
document repository. The information that is
necessary to process a document is stored in an
XML-format that is partially depicted in figure 3.
The left-hand side of the figure shows an excerpt of
a form, which students have to maintain during a
practical phase of their education.
In this example there is one anchor, namely the
word “Laufzettel” in the headline of the document.
When this anchor is found the upper left-hand corner
of the surrounding box of the anchor will serve as
origin for the positions of the fields in the document.
Then completion of the form can start. The
information for the relevant input fields is stored in
the XML-format file, which is partially depicted on
the right-hand side of figure 1. Here, the definition
of the data fields is shown in some detail. For each
field the distance to the anchor is given and
additional information like a field id and a fieldname
is also available. The element <FieldValue>
contains the corresponding value filled in the form
on the left-hand side. This value is determined at
runtime when the particular form template is
processed. The FABEGG system will try to read the
text at the specified position in the document. In a
first step the text read can be checked for syntactical
correctness. In this case the system can check the
text whether it coincides with the type given in the
<FieldType>-Element. The function that should be
used is specified in the <ValidateFunc>-Element
referring to a function called “Check4ValidName”
in our example.
The user interface consists of a camera unit,
speakers, and a display. The camera unit enables
easy document handling and fast document capture.
The interface consists of standard hardware only and
hence, such a user terminal could be placed in any
major administrative department. In a real
environment the speakers could be headphones due
to privacy considerations.
On the display the document image captured by
the camera is displayed, where . different colour
combinations can be chosen. The image processing
is done in real-time (Hennen, C., 2004). Moreover,
after the document has been processed by an OCR
software it can be read by using a speech synthesis
software allowing text output in different velocities.
While reading the text the system highlights the
word which is actually spoken by drawing a
coloured rectangle around it. This enforces auditive
understanding. Other features like reading word by
word or sentence by sentence are already included or
can easily be considered. Along with this features
the FABEGG system provides humans with visual
impairements, dyslexic people or people with
problems in reading comprehension with better
access to printed information. The system is also
able to process common electronic formats like PDF
or HTML. The overall process is scetched in figure
In order to process a form based document the
user hands over a digital image to the FABEGG
system, e.g. by taking a picture with the camera in
the FABEGG station or by using a scanner. Then,
FABEGG automatically determines the type of the
form and retrieves the matching template. Thus, it
knows which fields should be already filled e.g. by
an authority and which ones have to be completed
by the actual user. The regarding control information
is given with the help of the XML-document in the
repository. FABEGG guides the user through the
process of document completion. Therefore, it can
read any information contained in the form and
provide help information on the expected input for
the fields.
After all required fields of the form are filled it is
transformed into a representation that can be inserted
into a further workflow. In some contexts it is
necessary therefore to print the document for having
the possibility to sign the form. In other cases the
form can be transmitted electronically to the next
recipient. For this purpose FABEGG supports
common electronic formats like PDF or XML.
These documents may then be sent by E-mail.
Another possibility is transferring the information by
using public interfaces, e.g. based on HTTP or even
Web Services. Hence, FABEGG also achieves an
improvement by avoiding media disruption in
process chains.
Figure 1: XML document representation.
ICEIS 2007 - International Conference on Enterprise Information Systems
form completion
- syntactic validation
- semantic validation
(contextual information)
e.g. Zipcode
names -> Zipcode
document /
input devices
sample output formats
Figure 2: Detailed process for user client.
In this paper we have presented the FABEGG
system, which provides humans with reading
disabilities with much better access to information
systems. In particular, access to printed documents,
which are still important in business processes or in
government applications, is substantially improved.
We have shown how the form-handling component
allows inserting forms into the system and how to
retrieve them later on. By identifying the
corresponding process using the information in the
document template repository FABEGG integrates
paper documents into electronic process chains. As
electronic forms can be processed as well, FABEGG
also helps to prevent media disruption. Furthermore,
we can cope with documents where form fields have
already been filled with individual information in the
process before the user receives the form. This type
of communication is quite general in nowadays
business processes and cannot be replaced easily by
switching to an electronic version.
We will evaluate the FABEGG system together
with some local government authorities in Germany.
This will give further insight to needs of the
intended user groups. Furthermore, we consider to
implement FABEGG as a client / server application
thus giving more possibilities for the user front-end.
This will enable the users to have access to the
system from their home. Then they can also
maintain their documents in their personal document
management system.
Ballenger, W. L. 1979. Objective and alternatives for a
computer assisted instruction system for the visually
handicapped. SIGLASH Newsletter. 12, . 1979, 6-28.
Bundesrepublik Deutschland, 2006. Allgemeines
Gleichbehandlungsgesetz der Bundesrepublik
Edwards, A. D., 1988. The design of auditory interfaces
for visually disabled users. Proceedings of the SIGCHI
Conference on Human Factors in Computing Systems
(Washington, D.C., United States, May 15 - 19, 1988).
J. J. O'Hare, Ed. CHI '88. ACM Press, New York, NY,
European Commission, 2000. Gleichbehandlung
Behinderter in Beruf und Bildung: Richtlinie des Rates
2000/78/EG vom 27. November 2000, ABl. L 303
vom 2. Dezember 2000.
Hennen, C., 2004: Design and Implementation of
synchronous manipulation of a camera image with
spezial emphasis on ergonomics aspects. Diploma
thesis, Birkenfeld.
Muller, M. J., Wharton, C., McIver, W. J., and Laux, L.
1997. Toward an HCI research and practice agenda
based on human needs and social responsibility.
Proceedings of the SIGCHI Conference on Human
Factors in Computing Systems (Atlanta, Georgia,
United States, March 22 - 27, 1997). S. Pemberton,
Ed. CHI '97. ACM Press, New York, NY, 155-161.
Sun Microsystems, 2006: Java Accessibility API, Sun
United States of America, 1990. Public Law 101-336,
1990. Text of the Americans with Disabilities Act,
Public Law 336 of the 101st Congress, 1990.
W3C, 2006: Guidelines and resources from the W3C.