There is a powerful preoccupation at the European level in helping the people with such disorders;
that is why the EU Quality of Life and Management of Living Resources program has been developed. The OLP (Ortho-Logo-Paedia) project [1] for speech therapy has been started in 2002; the EU finances it and
it is a complex project, involving the Institute for Language and Speech Processing in Athens and seven other partners from the universitary and medical domains. The scope aims to accomplish a three –
modules system (OPTACIA, GRIFOS and TELEMACHOS) capable of interactively instructing the children suffering from dysarthria (difficulty in articulating words due to disease of the central nervous system).
The proposed interactive environment is a visual one and is adapted to the subjects' age (games, animations). The audio and video interface with the human subject will be the OPTACIA module, the GRIFOS
module will make pronunciation recognition and the computer aided instructing will be integrated in the third module – TELEMACHOS. Until now, the only public made results are a few articles about this
subject, while the research is in progress. 2.1. Priorities, directions, known objectivesThe priorities on the international level are represented by the developing information systems that will allow the elaboration of personalized therapeutical
paths. The following main directions are considered:
§ development of expert systems that personalizing the therapeutical guides to the child's evolution
§ evaluation of the motivation and progresses that the child's achieves
The most important objectives are determination of methods for the evaluation of speech impairments [2] where the ground truth data set is based on children between 2 years and 2 years and 11 months for
the English language. 2.2. State of the art, significant results Up to know, the only public available research are the scientifical articles [2] but research is still in
progress on the subject. An interesting project is STAR - Speech Training, Assessment, and Remediation [3], started in 2001, but is still in the developing phase. The members (AI. duPont Hospital for
Children and The University of Delaware) aim to build a system that would initially recognize phonemes and then sentences. This research group offers a voice generation system (ModelTalker) and other open
source applications for audio processing. On the international level, Speechviewer III developped by IBM that creates an interactive visual model of speech while users practice several speech aspects (e.g.
the sound voice or special aspects from current speech). The ICATIANI device developed by TLATOA Speech Processing Group, CENTIA Universidad de las Américas, Puebla Cholula, Pue., México uses sounds and
graphics in order to ensure the practice of Spanish mexican pronunciation (Kirschning, 2003). Each lesson explains sounds pronunciation using the facial expression with a particular accent on specifying
articulation points and the position of the lips. The system includes several animated faces, each of them showing the correct method of vocal pronunciation and providing feedback to the child answers. In
this case, if the child's pronunciation matches the system one, the child is rewarded by a smile or warned by a sad face otherwise. 2.3. Semnificatifs results and practical methods of application At the national level, little research has been conducted on the therapy of speech impairments, out of which mostly is focused on traditional areas such as voice recognition, voice synthesis and
voice authentication. We can mention the studies leaded by Prof. Iolanda Tobolcea at the Psychology and Education Science Department from "Al. I. Cuza" University, Iaºi. These studies have lead
to developing software for aided instructing that will provide feedback regarding the oral fluency. Although there are a lot of children with speech disorder, the methods used today in logopaedia are mostly
based on individual work with each child. The few existent computer assisted programs in Romania don't provide any feedback. At international level, there are software applications but quite
expensive(500$-1500$) and improper for the phonetic specific of Romanian language In conclusion, although present at the world level, the researches are not specific to the Romanian language and
at the national level all the researches, excepting those regarding speech disorders, are in incipient phases. With regards to the SoC systems (the solution we adopted for the mobile device), there are two main categories architectures:
§ SoC that are described in HDL languages (Verilog or VHDL) (for example Wishbone
System6800/01, System09 or C-NIT processor)
§ SoC that can be implemented using any kind of
FPGA, CPLD or ASIC and commercial SoC (the TriCore family from Infineon Gmbh. This SoC has the following interfaces built-in: Ethernet controller, USB 1.1 interface, MultiCAN interface, 3 UART
interfaces, IIC interface, 2 high speed Micro Link interfaces)
There is a lot of interest at the University of Suceava for SoC systems and research has been conducted and several results obtained for two PhD thesis having as director prof. dr. ing. Adrian Graur. 2.4. R&D organizations in the field
§ Institute for Language and Speech Processing Atena, University of Delaware, AI. duPont Hospital for
Children, Laboratoire Unformatique de Lille - Univ. des Sciences et Technologues de Lille, Institute Galilee - Univ. Paris Nord,
§ Universitatea „Alexandru Ioan Cua" Iasi – Facultatea de Psihologie, Universitatea Tehnica "Gh. Asachi" Iasi – Facultatea de Automatica si Calculatoare,
Universitatea "Stefan cel Mare" Suceava
2.5. Potential users
§ Children affected by speech impairments
§ Logopaed professors
|
