DEUTSCHER AKADEMISCHER AUSTAUSCHDIENST

 

Programm des Projektbezogenen Personenaustauschs (PPP) mit

Project Based Personnel Exchange Programme with

 

der SLOWAKISCHEN REPUBLIK

..........................................................................

Antragsformular für deutsche und slowakische Antragsteller/Application Form for Slovak and German Applicants

- bitte in deutscher oder englischer Sprache und unbedingt gemäß länderspezifischer Hinweise ausfüllen -

- Please complete in German (or English) and in accordance with the instructions given for the country concerned -

 

 

Neuantrag

New Application

x

Verlängerungsantrag

Application for Extension

r

 

 

 

2. Projektjahr/Year 2 of project

3. Projektjahr/Year 3 of project

r

r

 

 

1.Titel des gemeinsamen Projekts/

Title of joint project::

 

Audio-Speech Interface for Mobile Devices (ASIMD)

 

2. Fachblock/Subject block:

Communications

3. Fachrichtung/Field of specialisation:

Multimedia Processing

 

4. Slowakische Institution / Slovak institution:

x Hochschule/Institution of higher education

r Außeruniversitäre Forschungseinrichtung/Non-

university research institution

Slovak University of Technology, Faculty of Electrical Engineering and Information Technology

Institut/Name of institute:

 

Department of Telecommunications

Postfach/Strasse/P.O.Box/Street

Ilkovicova 3

PLZ/Ort/Postcode/Town:

81219 Bratislava

Telefon/Telephone no.:

+421-2-68279414

Telefax/Fax no.:

+421-2-68279601

E-Mail:

Gregor.Rozinaj@stuba.sk

Projektverantwortlicher/Project leader:

(Hochschullehrer bzw. leitender Wissenschaftler einer außeruniversitären Forschungseinrichtung/university professor or

academic in leading position in non-university institution, as appropriate)

Assoc. Prof. Gregor Rozinaj

Stellung/Current position:

Academic Teacher

 

5. Deutsche Institution / German institution:

x Hochschule/Institution of higher education

r Außeruniversitäre Forschungseinrichtung/ Non-

university research institution

Hochschule Heilbronn

Institut/Name of institute:

 

Faculty T1 – Mechanics and Electronics

Postfach/Strasse/P.O. Box/Street:

Max-Planck-Str. 39

PLZ/Ort/Postcode/Town:

D-74081 Heilbronn, Germany

Telefon/Telephone no.:

+49-7131-504-307

Telefax/Fax no.:

+49-7131-504-14-307-2

E-Mail:

gruhler@hs-heilbronn.de

Projektverantwortlicher/Project leader:

Prof. Dipl.- Phys. Gerhard Gruhler

Stellung/Current position:

Academic Teacher

 

6. Geplante Gesamtlaufzeit des beantragten Projekts/Proposed duration of project/

(Nur bei Neuanträgen auszufüllen/Only to be completed for new projects):

 

Beginn/Start: 01.01.2010

 

Ende/End: 31.12.2011

 

7. Förderung des Erstantrags/First Year Funding (Nur bei Verlängerungsanträgen auszufüllen/Only to be completed for applications for an extension):

 

Kalenderjahr/Year

 

Verbrauchte Mittel/

Funding expended

Bewilligte Mittel/Funding granted

 

 

DM/EUR

DM/EUR

 

DM/EUR

DM/EUR

 

8. Angaben zur Drittmittelförderung außer DAAD(Projekt-, Personal- und Reisemittel auf deutscher Seite Details of non-DAAD third party funding (Funding for projects, personnel and travel expenses):

 

Es besteht keine Drittmittelförderung/No third party funding                                                          x

Drittmittelförderung besteht oder wurde beantragt (Bei Bedarf zusätzliches Blatt beifügen)          r

Third party funding in existence or applied for (where necessary, please complete on a separate sheet)

Drittmittelgeber/Third party funding provider:

 

Mittelumfang/Amount:

 

Verwendungszweck/Purpose:

 

 

Laufzeit/Duration:

 

 

9.a) Frühere Förderung im Rahmen von PPP/Previous funding under PPP: x

 

Jahr/Year:

 

Slowakischer  Projektverantwortlicher/

Slovak  project leader:

 

Deutscher  Projektverantwortlicher/

German project leader:

 

Förderungslaufzeit/Duration of support:

 

Titel des gemeinsamen Projektes/

Title of joint project:

 

 

 

 

 

9.b) Wiederbewerbung nach Ablehnung/ Re-application following earlier rejection:x

Jahr/Year:

 

Slowakischer Projektverantwortlicher/

Slovak project leader:

 

Deutscher Projektverantwortlicher/

German project leader:

 

Beantragte Förderungslaufzeit/

Length of support applied for:

 

Titel des gemeinsamen Projektes/

Title of joint project:

 

 

 

 

 

 

 

 

 

 

 

12. Projektbeschreibung für den Neuantrag/Description of project for new applications

 

 

The aim of the project is to develop an optimized acoustic user interface for mobile devices suitable for the needs of the 50+ generation. Main objectives are:

·       to work out specifications for an optimized acoustical feedback of on-screen control elements

·       to implement synthesized acoustical feedback to touch-screens of mobile devices

·       to optimize multilingual speech output focussed on intelligibility for elderly and hearing-impaired users

·       to examine increase of performance of untrained users in comparison to standard devices

·       to optimize the database needed for speech-synthesis

 

 

The increasing complexity of modern personal electronic devices as well as the emergence of new types of these devices requires high flexibility of the potential users. Due to the demographic change in the population, the typical users tend to be older. There is already a market for mobile phones aimed at a relatively old group of users. These telephones often are limited to basic functions and feature large control elements and displays.

 

Furthermore a trend of convergence of several mobile devices can be observed. However, devices which can be used for more than one task tend to be more difficult to use than bespoke devices since either numerous control elements or a complex menu structure are required to make use of different functionalities. The usage of touch screens provides a possible solution for this challenge.  The benefit of a touch screen solution is that only buttons which are actually needed for the current functions have to be shown and use up space. Almost the entire footprint of the device is available as display area (for example for map display on a navigation device), while the same area can be used for input as well (i.e. address input on a navigation device or dialling on a mobile phone). However, a big disadvantage of virtual buttons is their lack of feedback. While a mechanical button provides the user with tactile and acoustical feedback, the user of a touch screen device has to rely on the visual feedback on the display to ensure that a button has been successfully pressed. Providing tactile feedback would require the usage of additional expensive actors, while loudspeakers usable for acoustical feedback are typically present on most mobile devices.

 

Synthesized speech output on the other hand makes it possible to read relatively long texts like e-mails to the user instead displaying it in either a very small font or with the need of scrolling.

The speech synthesis will be developed in Java as a diphone synthesis. The stress will be given to the optimization of the speech database. A novel algorithm for database compression based on sinusoidal parameterization of the diphones is planned in this application.

 

The current market for mobile phones aimed at the generation 50+ mainly consists of very simple devices with small displays and large control elements. This project's aim of combining the advantages of modern smartphone-type devices with a comprehensible interface would open this technology to new groups of people. The choice of Java platform for the implementation results in an easy implementation on various vendors devices.

 

The Digital Signal Processing Lab of the Institute of Applied research (IAF) at Heilbronn University is working since 2005 on the development of acoustic measurement strategies and systems for control elements in the human machine interface.

 

Furthermore, a rapid prototyping system for the acoustical feedback of control elements has been developed. This system enables human test subjects to individually design acoustic feedback sounds according to the subject's taste. The system has been originally designed to gather psycho-acoustically meaningful data to establish specifications for mechanical switches. The used synthesis algorithms however can be used to add meaningful auditory icons to graphical user interfaces as well as to facilitate the use of virtual buttons on a touch screen.

 

A test environment for subjective evaluation of the acoustic impression of control elements has been established. The test environment can be configured to perform numerous standard test procedures such as paired comparisons or magnitude estimation. The system can also be used for performance measurements in which the subject has to perform a task with varying user interfaces. The subjects speed and rate-of-error are observed.

 

The Slovak partner has been oriented to the research field of speech recognition and synthesis for approx. 15 years. Gregor Rozinaj, the leader of the Slovak research team has obtained several international patents in speech processing techniques. The partner in Slovakia provides a fully functional speech interface for mobile devices. It is based on diphone synthesis. It includes a talking-head application which is currently optimized for Slovak language.  However, the interface can be adapted to German and other languages with relatively little effort. Furthermore the Slovak partners provide a web-based tool for subjective evaluation of the quality of speech-synthesis.

 

German side has gathered know-how in sound-design for human machine interfaces. Furthermore, rapid-prototyping systems for usability tests and subjective evaluation exist in the German laboratory. The Slovak side also provides tools for subjective evaluation, but their main tool is a website for quick online survey of a relatively large number of subjects with limited possibilities regarding the control of the test environment and the modalities which can be simulated. However, the combination of online tests for quick pre-tests and the controlled laboratory environment in Heilbronn for in-depth evaluation of a prototype system increases the effectivity of subjective tests.

 

The Slovak side provides a highly developed and platform independent speech synthesizer for the output of long texts. Experience in transforming the synthesizer to other European languages (Italian, Greek and Romanian) promises a straightforward development of German speech synthesizer, too. The know-how on the German side regarding the perceived acoustic quality of control elements on the other hand optimizes the user input.

 

 

Except from the respective project leaders the work on both sides is done by PhD-students as well as students working on their Master or Bachelor thesis.

 

The growing demand for mobile devices (phones, navigation devices, ...) for the generation 50+ arises the need for comprehensible user interfaces. This project provides practically implementable solutions for this challenge. The implementation will be achieved using platform-independent JAVA programming.

 

While the teams in both countries achieved already advanced solutions in their specific fields, the aim of optimization of the acoustic user interface of mobile devices can only be achieved by cooperation and knowledge transfer. The German side gains broader knowledge especially in the field of speech synthesis. This fits into the general focus of the German team regarding improvements in the human-machine-interface. 

 

The co-operation should enhance the mobilities in between the both partners also on student level. For the next step it is planned to apply to EU grants in the field of the IKM projects.

           

 

 

13. Projektbeschreibung für den Verlängerungsantrag/Description of project for applications for an extension

 

 

 

 

14. Anlagen/Annexes

 

 

 

·      CV and publication list of the Slovak Project leader

 

 

 

 

 

 

 

 

 

15. Hinweis/Notes

 

Die Hinweise in der Ausschreibung und im Merkblatt zur Anfertigung des Antrags zum Projektbezogenen Personenaustausch habe ich zur Kenntnis genommen, insbesondere, daß mein Antrag aus formalen Gründen nicht bearbeitet werden kann, wenn der ausländische Partner nicht termingerecht seinen Antrag bei der Partnerorganisation vorgelegt hat. Auch ein in wesentlichen Teilen unvollständiger Antrag kann aus formalen Gründen abgelehnt werden.

I have taken account of the notes contained in the scholarship offer and the information sheet concerning the

completion of the application for funding under PPP, in particular, that my application will not be processed if the foreign partner’s application is not handed in by due date, and that incomplete applications may be rejected on formal grounds.

 

 

 

 

 

 

16. Datum/Date

 

Unterschrift des slowakischen Projektverantwortlichen/

Signature of Slovak Project Leader

 

 

 

 

 

17. Datum/Date

 

Unterschrift des deutschen Projektverantwortlichen/

Signature of  German Project Leader

 

 

 

 

 

 

18. Sichtvermerk der Verwaltungen der Hochschulen/Forschungseinrichtungen

 

Die Verwaltung erklärt sich bereit, im Falle einer Förderung des o.g. Projekts die Einzelabrechnung der Förderleistungen für alle Projektteilnehmer zu gewährleisten.

The management declares itself willing, should funding for the above project be granted, to provide separate accounts of assistance payments to all project participants.

 

 

 

 

 

Datum/Date

 

Unterschrift der Hochschule/Forschungseinrichtung/

Signature of university/research institution management

 

Siegel/Stamp

 

 

 

 

 

 

 

Datum/Date

 

Unterschrift der Hochschule/Forschungseinrichtung/

Signature of university/research institution management

 

Siegel/Stamp

 


Assoc. Prof. Gregor Rozinaj, Slovak project leader

 

Gregor Rozinaj received his M. Sc., Ph.D. and Assoc. Prof. degrees in telecommunications from the Slovak University of Technology (STU), Bratislava, Slovakia in 1981, 1990 and 1998, respectively. He has been a lecturer at STU since 1981. In 1992-1994 he worked on speech recognition at Alcatel in Stuttgart, Germany. From 1994 to 1996 he was with the University of Stuttgart, Germany. Since 1997 he has been Head of the DSP group at the STU, Bratislava. He is an author of 3 international patents on digital speech recognition and 1 patent on fast algorithms for DSP. Dr. Rozinaj is a member of IET. His main research interest is oriented to fast algorithms for DSP and speech processing.

 

Research Projects

Running Projects

 

Algorithms and Methods of Multimedia Signal Processing for Human Machine Interface

Rozinaj Gregor, VEGA 1/0718/09, (2009-2011)

 

Prosody and Modification of Speech

Rozinaj Gregor, APVV LPP-078-06, (2007-2010)

 

IQ Kiosk - Intelligent Terminal

Rozinaj Gregor, AV 4/0020/07, (2007-2009)

 

Finished Projects

 

Nonlinear Processing of Multimedia and Biomedical Signals in Telecommunications

Rozinaj Gregor, VEGA 1/3110/06, (2006-2008)

 

Measurement and Diagnostics of Speech Signal Prosody

Rozinaj Gregor, AV 123/06, (2006-2008)

 

Analysis and Processing of Low-level ECG Signals with Diagnostics in Practice

Rozinaj Gregor, VTP 18/2004 (2004-2006)

 

The Intelligent Speech Communication Interface, State Program of the Research and Development - Building of the Information Society

Rozinaj Gregor (university coordinator), 2003 SP 20 028 01 03 (2003-2006)

 

Audio, Video and Biomedical Signal Processing Using Non-Standard Forms of DSP Algorithms

Rozinaj Gregor VEGA 0147/2003 (2003-2006)

 

Virtual Reality of Multimedia Speech Synthesis

Rozinaj Gregor VTP 1003/2003 (2003-2006)

 

 

Selected Patents:

Rozinaj,G., Fesseler,P.:

Method and Circuit Arrangement for Detecting Speech in a Telephone Terminal from a Remote Speaker

U.S. Pat. Nr. US5818928, 6. Oct. 1998, Can. Pat. Nr. CA2187812, 6. Oct. 1998

 

Hörmann,Th., Rozinaj,G.:

Start-Endpunkt-Detektion zur Worterkennung,

(Detection of the start/end of words for word recognition)

Europ. Pat. Nr. EP0690436, 3. Jan. 1996.

 

Kopp,D., Rozinaj,G., Thierer,G.:

Pattern and Speech Recognition using Gradient-Based Dynamical Reduction of DP Matching Search Area

U.S. Pat. Nr. US5737722, 7. Apr. 1998


 

Selected publications:

Marcek, A.; Kotuliaková, J.; Rozinaj, G.:New Approach Of Fast ICT And MICT Algorithms Development IEEE ICECS '96, Rodos, Greece, Oct. 13-16, 1996, pp. 744-747  (citations 8/6)

 

Rozinaj, G.; Polec, J.; Kotuliaková, J.; Podhradský , P.; Marcek, A.; Marchevský, S.:Digital Signal Processing II (in Slovak) FABER Bratislava 1997, ISBN 80-967503-5-6, 155 pp. (citations 10/8)

 

Kacur,J.; Rozinaj,G.; Herrera,S.:Word Boundary Detection Based on Global Variatiation Maesuring of CLPC Vectors Using Cepstral Matrices, International Conference on Telecommunications ICT'2000, Acapulco, Mexico, May 22-25, 2000, pp. 691-695

 

Rozinaj, G.: Frequency Acquisition in a Signal with Few Sinusoids Using Effective Fourier Transform  Turkish Journal of Physics, Vol. 20, Num. 4, 1996, ISSN 1300-0101, pp.324-330

 

Macho, D.; Nadeu, C.; Jancovic, P.; Rozinaj, G.; Hernando,J.: Comparison of Time & Frequency Filtering and Cepstral-Time Matrix Approaches in ASR In: EUROSPEECH'99, Budapest, Hungary, 5. Sept. , 1999, pp. 77-80 (citations 6/5)

 

Kaèur, J., Rozinaj, G.: Word Boundary Detection in Stationary Noises Using Cepstral Matrices, Journal of Electrical Engineering, Volume 54, 2003

 

Kaèur, J., Rozinaj, G., Herrera-Garcia, S.: Speech Signal Detection in the Noisy Environment Using Neural Networks and Cepstral Matrices,  Journal of Electrical Engineering, Volume 55, 2004

 

Turi Nagy, M., Rozinaj, G.: An Analysis/Synthesis System of Audio Signal with Utilization of an SN model, In: Radioengineering Vol.13, pp.51-57,  December 2004

 

Vajaš, M., Rozinaj, G.: Facial Animation During Speech Performance, IASTED International Conference on Communications, Internet and Information Technology (CIIT 2004),November 22-24, 2004, St. Thomas, USA

 

Juhár, Jozef - Ondáš, Stanislav - Èižmár, Anton - Rusko, Milan - Rozinaj, Gregor - Jarina, Roman:
Development of Slovak GALAXY/VoiceXML Based Spoken Language Dialogue system to Retrieve Information from the Internet, In: INTERSPEECH 2006 : 9th International Conference on Spoken Language Processing. Pittsburgh, USA, 17.-21.9.2006. - CD-Rom

 

 

Juhar, Jozef - Cizmar, Anton - Rusko, Milan - Trnka, Marian - Rozinaj, Gregor – Jarina, Roman: Voice Operated Information System in Slovak, In: Computing and Informatics, Volume 26, 2007, Number 6, ISSN 1335-9150

 

Talafová, Renata - Rozinaj, Gregor – Cepko Jozef - Vrabec, Ján: Multimedia SMS Reading in Mobile Phone, In: INTERNATIONAL JOURNAL of MATHEMATICS AND COMPUTERS IN SIMULATION, Issue 1, Volume 1, 2007, ISSN 1998-0159

 

Oravec, Milos - Beszedes, Marian - Rozinaj, Gregor: Detection and Recognition of Human Faces and Facial Features In: Prasad, Bhanu; Prasanna, S.R.M. (Eds.): Speech, Audio, Image and Biomedical Signal Processing using Neural Networks, Springer Verlag, 2008, X, 420 p., Hardcover, ISBN: 978-3-540-75397-1, pp.: 283-305

 

Rozinaj, Gregor - Vrabec, Ján - Cepko Jozef - Talafová, Renata: Terminals for the Smart Information Retrieval

In: Ismail Khalil Ibrahim (Ed.): Handbook of Research on Mobile Multimedia, Second Edition, Chapter XIX, IGI Global Publishing, 2008, ISBN: 978-1-60566-046-2, 451 pp.