Automatic Generation of LIBRAS Signs by Graphic Symbols of

SignWriting

Carlos E. A. Iatskiu, Laura S. Garc

ıa and Rafael P. Canteri

Federal University of Paran

a, Curitiba, PR, Brazil

Keywords:

Deaf People, Deaf Community, LIBRAS, SignWriting.

Abstract:

The Brazilian Sign Language is the natural language used by Deaf people in Brazil to communicate between

themselves and with the society, as well as it is part of culture and tradition. Providing access to communi-

cation, information and knowledge (creation) for the Deaf community are just some of the motivations for

Brazilian Sign Language writing record. This paper shows some hypotheses for the low usage of computa-

tional tools for recording sign languages and proposes a new way to generate the graphic records in Brazilian

Sign Language through the SignWriting System for assisting the Deaf individuals in the exercise of their full

citizenship.

1 INTRODUCTION

Despite the latest and emerging worldwide social

transformations, people with disabilities still suffer

from several problems such as prejudice and inacces-

sibility, among others. Gradually and slowly, transfor-

mations in all knowledge areas have started to solve

some of these problems. Deaf communities were for-

merly excluded from society due to the fact of their

consideration as not “normal” (clinically originated

concept) individuals. Reports show the extreme pro-

cedures to which Deaf people used to be submitted

in order to “remove their abnormality”. Nowadays

things have changed. Nevertheless, the above ques-

tions are far from having been attended, and social

errors and misconceptions about handicapped needs

goes on generating (new forms of) exclusion (Fernan-

des, 2007).

Deaf communities (born Deaf people with proper

culture in opposition to people with acquired deaf-

ness) in the world still suffer from the lack of oppor-

tunities and, in this sense, tools that can help them in

communicating with each other and with other mem-

bers of society are necessary. Plain access to infor-

mation is one of the research challenges referring to

Deaf people, and since information systems help to

meet this goal, they are critically important. However,

these systems usually impose obstacles to access and

use. One of the obstacles is that the human computer

interaction is not natural for these populations, be-

cause communication via interface is not provided in

sign language - in this case LIBRAS (Brazilian Sign

Language), considered by linguistic experts the ﬁrst

language of Deaf people, even when it is not chrono-

logically the ﬁrst one.

From the observation of this reality the need for

computational tools that can help Deaf people in com-

munication and access to information becomes clear.

In conjunction with all the needs of this communities,

there is the difﬁculty they face with the graphic record

of LIBRAS, namely, the problem determined by the

lack of a standard writing system and of the corre-

sponding practice of its use by Deaf communities.

No language can survive without its writing system,

because languages evolve along time. Also, without

a proper record, part of them is lost. Deaf commu-

nities in Brazil use LIBRAS for communication be-

tween their members, but every time they need to use

some record of it, they are restricted to use written

Portuguese, their second language.

Human beings have different volatile ways of

communication, which passed from people to peo-

ple, with the inherent risk of their meaning modi-

ﬁcation. Writing systems remain intact. From this

the need for the written registration of sign languages

arises. According to Dulcineia Azeredo (Azevedo,

2010), “The importance of writing for history and for

preservation of the records comes from the fact that

it allows for the storage and propagation of informa-

tion, not only between individuals, but also through

generations. Through the writing, it is possible to

put the present man in contact with the man from the

Andrade Iatskiu C., Sánchez García L. and dos Passos Canteri R..

Automatic Generation of LIBRAS Signs by Graphic Symbols of SignWriting.

DOI: 10.5220/0005372200710078

In Proceedings of the 17th International Conference on Enterprise Information Systems (ICEIS-2015), pages 71-78

ISBN: 978-989-758-098-7

 2015 SCITEPRESS (Science and Technology Publications, Lda.)

past, rescuing facts to be discussed in a new perspec-

tive. This will certainly have beneﬁts for human exis-

tence.”

In this way, Deaf people who know a sign lan-

guage should be able to represent and record it

through writing. Additionally, one useful type of

tool to assist the Deaf in the process of maintaining

their culture and recording the knowledge produced

by them is a sign writing system. Many different writ-

ing systems for sign languages emerged trying to re-

solve sign representation, but they have produced a

new difﬁculty for Deaf people, since they rarely know

which system is more appropriate or, even worse, they

are not able to take advantage of the system because

of the difﬁculty of interacting with it.

The most well known sign language writing sys-

tem is SignWriting, a visual notation whose elements

can be easily mapped to sign languages phonological

components. Its elements allow for any sign language

to be visually represented, ie, the writing system al-

lows to represent the sign phonological structures in

graphical form.

SignWriting emerged as an evolution of an

spelling system for representing ballet movements

created by Valerie Sutton, around the 70s, at the Uni-

versity of Copenhagen in Denmark, the DanceWriting

(Sutton, 2005).

Writing systems show interesting alternatives to

assist the written record of signs. SignWriting is a

complete writing system with all the potentiality to

be used in Brazilian Sign Language (LIBRAS) to fa-

cilitate the description and the understanding of sign

language records. The writing system presents pos-

sibilities to express the majority of the sign language

resources, as well as its visual spatial modulators in-

corporated in the signs in context. These character-

istics determined our choice of SignWriting for the

development of the tool in question (Stumpf, 2005)..

In Brazil, LIBRAS has been used since 1996 in

Deaf education and in research by Marianne Rossi

Stumpf, who was the ﬁrst Deaf to write LIBRAS’

signs. Then, she started researching about sign writ-

ing with Professors Ant

onio Carlos da Rocha Costa

and Marcia Campos at the Pontiﬁcial Catholic Uni-

versity in the state of Rio Grande do Sul. From there,

Marianne and Antnio translated a children’s book “A

girl is called Kauana” from Portuguese to SignWrit-

ing. This was the ﬁrst text written in LIBRAS in

Brazil (Stumpf, 2000).

Despite this ﬁrst interest, the use of SignWriting

is still very restricted, but many authors defend that

its use could assist the target population in expand-

ing and documenting their language, LIBRAS. This

writing system is still not known by the Deaf com-

munities. One of the arguments raised by the Deaf

is that SignWriting is not used mainly because most

of information and knowledge legacy is recorded in

written oral languages, as Portuguese, and learning

SignWriting is in fact an indirection towards written

Portuguese domain. Additionally, there are few tools

that support LIBRAS recording in SignWriting.

Related literature showed some editors and/or

translators that support SignWriting, but these tools

have not obtained acceptance (considered as usage

practice) of the Deaf communities in Brazil or around

the world. From the study of these tools, we have

built a complementary hypotheses, namely that this

aversion may have been caused by inadequate in-

terfaces and especially by the adoption of incorrect

paradigms, which demand high time to generate a

single sign, making it extremely difﬁcult to translate

extensive text from some country’s oral language to

SignWriting thus demotivating handouts, books and

others translation efforts.

In this context, the development of technological

artifacts with appropriate interfaces and interaction

models built on paradigms that encourage the use and

the legitimacy of SignWriting as a writing system for

LIBRAS appears as a sound research route.

In Section 2 describes a Review of Literature

(writing systems for Sign Languages and SignWrit-

ing, The Computational Architecture to Support the

Social Inclusion of Deaf Communities in Brazil and

Considerations of Computational Description Model

of LIBRAS), Section 3 presentes The Low (or NO)

Use of Computational Tools to Support SignWrit-

ing by Brazilian Deaf Communitty, Section 4 shows

Automatic Generation of LIBRAS Sign by Graphic

Symbols of SignWriting. Finally, Section 5 discusses

a Conclusion and Future Works.

2 REVIEW OF LITERATURE

In this section the writing systems for Sign Languages

and SignWriting (Stumpf, 2005), the Architecture

(Garc

ıa et al., 2013) (Garc

ıa et al., 2010a) (Garc

ıa

et al., 2010b) and the Phonological Model (Antunes,

2011) for adequate LIBRAS treatment, cornerstones

for the present WEB service, are presented.

2.1 Writing Systems for Sign

Languages and SignWriting

As it was already mentioned, there are several writ-

ing systems for sign languages around the world that

address the graphic representation of signs. However,

ICEIS2015-17thInternationalConferenceonEnterpriseInformationSystems

this is seen as an additional problem for Deaf commu-

nities that, unable to know which is the most adequate

system to use, appeal to the written form of the oral

language of their country, not without signiﬁcative ef-

fort. The present section shows some of the available

writing systems for sign languages, focusing on Sign-

Writing, which is considered the most complete one.

• Stokoe’s System: Stokoe (Stokoe, 1978) cre-

ated a notation to represent sign languages based

on their observed parameters: hand conﬁguration

(10), location (12 positions), action movements

(22) and hand orientation (4). He was also the

forerunner of the sign languages linguistic char-

acter legitimation. The goal of his notation was

to respond to the needs off the registration and

the study of sign languages. The use of the writ-

ten code by the Deafs themselves was not one of

his purposes. Stokoes System was the ﬁrst nota-

tion to represent the “phonological” (by the level

of equivalence to oral languages analysis) compo-

nents of the American Sign Language by graphic

symbols. The original notation was composed by

55 symbols, but afterwards, researchers modiﬁed

this set as a consequence of the productivity of

living languages (Stokoe, 1978).

• Franc¸ois Neve’s Notation: Franc¸ois Neve, a re-

searcher at the University of Li

ege (1996), ex-

panded the Stokoe’s notation complemented it.

The Franc¸ois Neve’s notation made the enumer-

ation and the computational treatment of the signs

possible. The written representation is done via

columns from the top to the bottom (in a single

column when the dominant hand signalizes or in

two columns for both hands) (Stumpf, 2005). The

phonological aspects speciﬁed up to that moment

were considered both in Stokoe’s and in Neve’s

notations. Both of them also had the primary el-

ements needed to represent non-manual expres-

sions (face, body and head, among others), sig-

niﬁcant components of sign languages.

• HamNoSys: Hamburg Sign Language Nota-

tion System (HamNoSys) (Prillwitz et al., 1989)

(R. Elliott and Parsons, 2001) is a phonological

linear transcription system with a larger scope

than Stokoes. This system deﬁnes about 200 sym-

bols for representation of parameters like hand

conﬁguration, orientation, locations in the head

and in the body and also the movements through

iconic representations easily perceived and rec-

ognized. This model also brought, as an inno-

vation, the inclusion of the non-manual expres-

sions. A computational system for sign transcrip-

tion was developed based on HamNoSys model,

and it later helped to propose tools for Deaf peo-

ple. The SiGML (Signing Gesture Markup Lan-

guage) (Prillwitz et al., 1989) is this transcription

model, built by means of XML (the eXtensible

Markup Language) and a ﬂexible alternative for

representing sign languages signs constituents in

a computational way.

2.1.1 SignWriting

SignWriting is a graphic written notation of sign lan-

guages. Its components allow for any sign language

to be represent visually, or, in other words, the writ-

ing system allows for representing the phonological

structure of sign languages in a graphic manner. Fig-

ure 1 shows examples of SignWriting.

Figure 1: Examples of SignWriting (Stumpf, 2005).

This writing system was developed by Valerie Sut-

ton in 1981 as an extension of a dance movement no-

tation system (Sutton, 2005), the Sutton Movement

Writing & Shorthand, that aimed at being able to rep-

resent, in a systematic way, any movement of dance,

sports, physiology among others.

In Capovilla & Raphael work (2001)(Capovilla

and Raphael, 2001), it is possible to perceive that

SignWriting has a fundamental role in sign languages

in general, since it helps the Deaf people to under-

stand efﬁciently each detail of the sign composition

by means of the written correspondent. Parameters

as location, movement, orientation, facial expressions

can be easily visualized in SignWriting.

SignWriting has a structure composed of four ba-

sic elements: hands, movements, facial expressions

and body that other writing systems do not include.

The elements for hand representation are the

hands conﬁguration, their orientation and movement.

These parameters enable the distinction of types of

contact made between one or both hands and the rest

of the body. The system has about 900 symbols,

which provide more accuracy to the representation of

signs (Stumpf, 2005). Facial expressions and body

movements are indispensable for the representation

of signs. Table 1 presents groups of classiﬁcation of

SignWriting symbols.

AutomaticGenerationofLIBRASSignsbyGraphicSymbolsofSignWriting

Figure 2: Integration Model of the proposed HCI Architecture with developed and in progress modules and bases. (Garc

ıa

et al., 2010a) (Garc

ıa et al., 2010b).

Table 1: Table with groups of SignWriting Symbols

(Stumpf, 2000).

Orientation and Position of Hands

Contact Types

Hands Conﬁguration

Finger Movements

Arms Movements and appointments (straight, curved, ...)

Facial Expressions

Location of Head Symbols

Head Movements

Orientations of the look

Body Movements

Symbols of Grammatical Points

Dynamic of the Movements

Even being SignWriting a graphic code for signs

writing, some important parameters such as velocity,

frequency, sequentiality marks, among others, are not

available. In socially situated use, these are perceived

in a natural and unconscious form by sign language

users. Nevertheless, for adequate computational treat-

ment, a more detailed sign phonological components

description is needed.

Further differential of SignWriting to other sim-

ilar systems refers to the fact that this system is the

only one that presents clearly characteristics related

to simultaneity and that utilizes a precise spacial de-

scription.

The writing systems presented here are seen as in-

teresting alternatives for helping in signs registration.

SignWriting, mainly, is a proper written system with

all the potentiality to be used in LIBRAS, as discussed

above. This justiﬁes its choice as the basis for the web

service of translating our research group phonological

model to the signs writing system in question.

2.2 The Computational Architecture to

Support the Social Inclusion of Deaf

Communities in Brazil

The architecture proposed by our research group can

be described trough two main elements: the ab-

stract four general layers hierarchy and the integra-

tion model, which shows all the modules and their in-

terrelations that give support for the execution of the

abstract architecture. Figure 2 shows the Integration

Model of the Architecture.

The four level hierarchy is show in Figure 4. The

surface layer is responsible for providing adequate ap-

plications, mainly in the axes of i) giving proper sup-

port to Deaf natural language (i.e. Sign Language)

acquisition and registration (as occurs traditionally in

the written codes of any oral language); ii) support-

ing teaching-learning processes of LIBRAS itself and

of written Portuguese as their second language (i.e.

literacy); iii) supporting knowledge acquisition of ev-

ery other area, considering the hypothesis of the Sign

Language acting as the mother language for the Deaf

in any interpretation process.

The needed services refer specially to those asso-

ciated to the language itself, starting by dictionaries,

lexicon and translators. Although being themselves

applications, they are critic as tools as well, allowing

for plain applications for Deaf communities. The in-

terface between the services and the internal APIs has

as its main function providing correct frameworks for

both uppermost layers from sound different speciali-

ties bases Finally, the internal level is responsible for

the Computer Science sub-areas knowledge and tech-

nology needed to support the several tools and appli-

cations.

After presenting the overall architecture, we pro-

ceed to describe the integrating model. This repre-

sentation exposes all the modules and bases involved,

and their interrelations. Figure 3 shows it, together

ICEIS2015-17thInternationalConferenceonEnterpriseInformationSystems

Figure 3: Abstraction of the proposed HCI Architecture to

promote inclusion and citizenship for Deaf communities.

(Garc

ıa et al., 2010a) (Garc

ıa et al., 2010b).

with the modules instantiated (developed/in progress)

up to the present moment.

The work described in the present paper - the

WEB service for the interpretation of the sign de-

scriptions by means of the computational model of

LIBRAS phonology to SignWriting symbols - is situ-

ated within the described architecture as indicated in

Figure 3, as part of the SignWriting Generator, since

it receives the computational phonological descrip-

tion of a sign components and produces, for the user,

the graphic representation components of the sign in

SignWriting.

2.3 Considerations on the LIBRAS

Phonology Description

Computational Model

Other participant of the research group in which the

present work is situated developed a computational

representation model based on a compilation and

adaptation of phonological models that aggregates a

high degree of details for sign description. Such ﬁ-

nesse in describing the signs render the model pre-

sented adequate for computational treatment of SL.

Figure 4 shows an adaptation of the basic structure of

the model with some of its main parameters, one can

see a Sign is composed by a Hold, a Non-Manual Ex-

pression and a Movement (that may be combined in

different levels to represent simultaneity and sequen-

tiality). Each element, by turn, is the roots of the tree

that expands down to the leaf level, where the actual

values of the required parameters are enumerated (e.g.

the element of quality may be deﬁned by the exten-

sion, tension, velocity and other temporal aspects of

the movement). The proposed structure (its use and

Figure 4: Conceptual Structure of Sign Computational Rep-

resentation (Antunes, 2011).

the degree of computational details of signs) is impor-

tant to solve a series of problems thus far encountered

in SL recognition, building signs (3D), written signs,

among others.

The proposed model is a compilation of several

phonological models existing in the literature, and it

extends these models in various aspects: it allows for

simultaneity, sequentiality; non-manual expressions

etc. The model has a high degree of detailed param-

eters and values that could be attributed to such pa-

rameters. Additionally, it allows for adoption of new

expressions, parameters and values, which is a pow-

erful feature in the model to support new signs in SL.

The model is critical as a proper support for image

recognition, translation, 3D avatars generation and so

on. Additionally, the proposed model includes the

non-manual expressions, that may be used in the in-

tensity, sentence formation, semantics and in the sin-

gular characterization of a sign. This singular charac-

terization often occurs in disguised signs. These signs

aggregate facial expressions. The proposed model is

an important element of the conceptual HCI architec-

ture previously cited that takes several aspects of Deaf

needs into consideration.

Figure 5 shows the sign of Brazilian Sign Lan-

guage that represents “tree”. In this sign, the

dominant-hand takes the form of the speciﬁc conﬁgu-

ration 56 used at the point of articulation. Note the

need contact attribute, because the signal is articu-

lated on the back of the non-dominant hand contact

with the elbow. Still to make the sign it is neces-

sary to use the local movement of rotation (counter-

clockwise). Table 2 describes this sign by the Com-

putational Description Model of the Phonology of LI-

BRAS.

Such model is appropriate, for example, to aid

Computer Vision (CV) to better train algorithms for

sign recognition, with the correct set of representa-

tive signs and descriptions. Most CV studies rely on

the use of gloves for recognition. A glove limits the

natural signing of the SL and systems of this kind do

not consider, in most cases, very important aspects as

AutomaticGenerationofLIBRASSignsbyGraphicSymbolsofSignWriting

Figure 5: Sign Tree “

arvore” (Antunes, 2011).

Table 2: Table with the XML Description of the Movement

of the Sign Yellow “amarelo” in LIBRAS.

<hand >dorse </hand >

</location >

<palm hand=“vertical”>upward, forward </palm >

<ﬂoorParallel>hand base </ﬂoorParallel>

</orientation>

</dominantHand>

<spatialRelationship loca-

tion=“stomach”i=“shoulder” >

</space >

<trunk >elbow </trunk >

</location >

<palm hand=“vertical”>upward, forward </palm >

<ﬂoorParallel>hand base </ﬂoorParallel>

</orientation >

</nonDdominantHand>

</hold >

</sign >

non- manual expressions, locations, among others.

To assist CV in the recognition process, the cre-

ation of an adequate sign base (not random or iso-

lated) with descriptions of the signs according to

phonology is important. Thus, CV will receive a rep-

resentative sets of signs and their recognition will de-

termine elements of phonology. In this way it is pos-

sible to retrospectively identify any other signs not

present at the base, by recognition of their parame-

ters. Thus, it is possible to create tools that allow

for searching signs by the signs’ phonological com-

ponents and will support translation process.

This model is a core feature in the work in ques-

tion, since from its rules for sign components descrip-

tion in a high level of detail, it permits the automatic

generation of the graphic symbols.

3 THE LOW (OR NO) USE OF

COMPUTATIONAL TOOLS TO

SUPPORT SIGNWRITING BY

BRAZILIAN DEAF

COMMUNITTY

Some results found represent the state of art for the

development of information systems for members of

Deaf communities in Brazil and worldwide. There are

some initiatives that aim to assist these communities

by the learning of sign languages and registering of

language through SignWriting.

Recent developed tools, although with the purpose

of helping the Deaf, are produced to a very speciﬁc

audience (those Deafs that are able to read and write

the ofﬁcial language of their country) and in this way,

they are not adopted in everyday activity by Deaf peo-

ple as communication resources, interaction or learn-

ing. Many of these artifacts do not attend in general

to the needs of Deaf people, among which reasons we

can mention language issues (i.g. the information is

presented mostly in the written language of the coun-

try, assuming that all the Deaf can read it) and inter-

action (i.g. most of the tools for users to interact with

the interface are not mediated by LIBRAS).

By analysing related work, it has been concluded

that most available tools have several problems which

can be inferred by the low usage reported. One of the

common problems is that, in general, the use of sym-

bols trees to edit a sign is a paradigm that should be

abandoned, because in order to edit a sign in a soft-

ware that uses this technique, it is necessary to drag

primitive by primitive and positioning it in the corre-

sponding place. This takes a long time, since the num-

ber of existing primitive symbols is extremely high,

being, then, unsuitable to write a long texts.

Another point that causes this low use of tools by

the community is the poor interfaces quality, some of

them still being developed in MS-DOS environment,

together with a high number of accessibility prob-

lems, hardening the user-tool interaction.

Also, the development of a new way to generate

the graphical form of the signs of LIBRAS on Sign-

Writing tools is necessary, ﬁrstly in order to provide

the Deaf community with viable edition processes to

allow for information accessing and knowledge cre-

ICEIS2015-17thInternationalConferenceonEnterpriseInformationSystems

ation and, secondly, to written the components that

will be the base for LIBRAS sign generation sign gen-

eration to provide Human-Computer interfaces writ-

ten outputs.

4 AUTOMATIC GENERATION OF

LIBRAS SIGNS BY GRAPHIC

SYMBOLS OF SIGNWRITING

This automatic generation is situated within the archi-

tecture as described in Figure 3. It can be used for the

automatic interpretation of XML to the SignWriting

graphic symbols, and, also, for intelligent search for

graphic symbols.

The system reads the LIBRAS sign description

in the Phonological Model in XML and provides as

output the graphic symbols of SignWriting that corre-

spond to the input sign.

Figure 6 shows the automatic generation process.

The main difference in relation to other tools avail-

able comes from the fact that it does not make a con-

version from a sign to a unique pre-existent SignWrit-

ing correspondent symbol sequence but, instead, con-

verts each piece of XML describing the primitives

that compose the sign (dominant hand, movement,...)

to the correspondent SignWriting sign component.

Figure 6: Abstraction of the proposed system (Iatskiu,

2014).

Figure 7 displays the LIBRAS graphic sign for the

Portuguese word “pai” (father). Table I should be ob-

served in order to understand the way in which the

system interprets each piece of XML code and how it

forms the written symbols. It shows the movement in

LIBRAS Phonological Model of Figure 4.

Figure 7: Graphic sign of “pai” father in SignWriting.

As we can observe in Table 3, the service reads

each described XML chunk in an independent man-

ner for the SignWriting symbols generation. In this

way, every existent (or yet non-existent, if represented

in the Model) LIBRAS sign can be generated from its

primitives description. The systems databases stor-

age every elementary component described by the LI-

BRAS Phonological Model, instead of the correspon-

dence sign-SignWriting symbols like other available

tools do.

Table 3: Table with the XML description of the movement

of the sign “pai” in LIBRAS.

<type>

<contour>rectilinear</contour>

<unidirectional>sideways</unidirectional>

<contact>touch</contact>

</type>

<velocity>normal</velocity>

</quality>

<frequency>double</frequency>

</dominantHand>

</movement>

The tool is relatively simple, it shows the user the

structure of the descriptive XML for him to enter the

LIBRAS signal information, which can also have the

clean interface shown in Figure 9, if the XML is ready

for automatic conversion. Thus, the system reads the

user input, corrects any mistakes and eliminates not

required information saving each description in a vec-

tor.

After this stage, the main vector is separated into

7 other vectors respectively connected to each Sign-

Writing primitive: dominant-hand conﬁguration, lo-

cal movement of dominant-hand, non-dominant hand

conﬁguration, local movement of non-dominant hand,

no manual expression, movement of dominant-hand

and non-dominant hand. A scan occurs to verify if

the information is primitive, because in a sign it is

not necessary to have all of these primitives and each

primitive does not need to have all the descriptions

recorded.

The ﬁnal stage of the process is to check if the

vector has the description of the primitive. If so, sev-

eral searches in the database are performed to ﬁnd

some primitive that has the similar characteristics to

the one inserted into the XML and if it ﬁnds it the

AutomaticGenerationofLIBRASSignsbyGraphicSymbolsofSignWriting

Service shows to the user the primitives forming the

corresponding signal. Figure 9, shows a screen of the

tool for Automatic Generation of LIBRAS Signs by

Graphic Symbols of SignWriting.

This makes the system’s managing possible and

innovative, since along the language evolution in

course, it will not be necessary to enter each new sign

created, because its elementary components will be

already storage.

The second functionality of the service presented

here acts in a similar way as other available tools do,

with the difference that it uses as input the same XML

code with the sign description in the ﬁrst capacity or

use of the service described above. The output for the

user consists on the set of all the signs that have that

chunk as a component. The service is all being devel-

oped for the WEB platform in order to be available

for Deaf communities. It has been declared especially

useful as a tool for LIBRAS learning. Also, it will be

integrated to the whole working architecture within a

non-distant future.

5 CONCLUSIONS AND FUTURE

WORK

The need of any natural language writing is critical,

so that the language is not lost along time. Deaf com-

munities needs are clear, mainly concerning access

to information without the country’s ofﬁcial language

indirection. In this sense, the work presented here po-

tentiality represents a step towards solving a relevant

technical problem, also having clear future social re-

sults. The service presented here is a relevant part of

an alternative output to 3D LIBRAS avatars within the

Portuguese - LIBRAS translator module under devel-

opment.

The main question to be treated in future work is

the generation of SignWriting symbols of a certain

to down, from left to right). This will be the last step

after the conversion of the written Portuguese ver-

sion into the corresponding XML LIBRAS descrip-

tion by the Phonological Model built within the re-

search group.

This capacity will be one of the bases for au-

tomatic translation from written Portuguese to LI-

BRAS, giving Brazilian Deaf communities the pos-

sibility of having access (in graphic written form) to

books, classroom notes, ﬁlms and any other academic

or cultural content and, last but not least, to knowl-

edge creation in their natural language.

ACKNOWLEDGEMENTS

We would like to thank the students and teachers that

collaborated to this research. This work is funded

by Coordenac¸

ao de Aperfeic¸oamento de Pessoal de

ıvel Superior - CAPES and Fundac¸

ao Arauc

aria.

REFERENCES

Antunes, D. R. (2011). Um modelo da descric¸

ao computa-

cional da fonologia da l

ıngua de sinais brasileira. In

Universidade Federal do Paran

a. UFPR - Curitiba.

Azevedo, D. (2010). A import

ancia da escrita na sociedade

contempor

anea. Universidade Est

acio de S

Capovilla, F. and Raphael, W. (2001). Dicion

ario En-

ciclopdico Ilustrado Tril

ıngue da L

ıngua de Sinais

Brasileira: Sinais de A L. Editora da Universidade

de S

ao Paulo, S

ao Paulo, i edition.

Fernandes, S. F. (2007). Avaliacc

ao em l

ıngua portuguesa

para alunos surdos: algumas considerac¸

oes. Universi-

dade Federal do Paran

Garc

ıa, L., aes, C. G., ands D. Trindade, D. A., and da Silva,

A. (2010a). Structure of the brazilian sign language

(libras) for computational tools: Citizenship and so-

cial inclusion. In World Summit on the Knowledge

Society. RCIS.

Garc

ıa, L., aes, C. G., ands D. Trindade, D. A., da Silva, A.,

and Junior, A. M. (2010b). Techonological artifacts

for social inclusion: Structure of the brazilian sign

language (libras)- gestures for citizenship. In IADIS

International Conference on WWW/Internet.

Garc

ıa, L., aes, C. G., Antunes, D., and Fernandes, S.

(2013). Hci architecture for deaf communities cultural

inclusion and citizenship. In ICEIS.

Iatskiu, C. (2014). Servic¸o WEB de Intepretac¸

ao do Modelo

Fonolgico Computacional da Libras para os S

ımbolos

aﬁcos do SignWriting. Universidade Federal do

Paran

a, Curitiba.

Prillwitz, S., Leven, R., Zienert, H., and Hanke, T. (1989).

Hamnosys version 2.0: Hamburg notation system for

sign languages: An introductory guide. In Interna-

tional Studies on Sign Language and the Communica-

tion of the Deaf.

R. Elliott, J. G. and Parsons, J. K. (2001). Sigml deﬁnition.

In In Technical report working document.

Stokoe, W. C. (1978). Sign language structure. In Silver

Spring: Linstok Press. Linstok Press.

Stumpf, M. R. (2000). L

ıngua de sinais: escrita dos sur-

dos na internet. In V Congresso Ibero americano de

Inform

atica Educativa.

Stumpf, M. R. (2005). Aprendizagem de Escrita de L

ıngua

de Sinais pelo Sistema SignWriting: L

ınguas de Sinais

no Papel e no Computador. Universidade Federal do

Rio Grande do Sul, Porto Alegre.

Sutton, V. (2005). Lic¸

oes sobre signwriting: Um sistema de

escrita para l

ıngua de sinais. In Traduc¸

ao de Marianne

Rossi Stumpf.

ICEIS2015-17thInternationalConferenceonEnterpriseInformationSystems