Tel : 501 077 373
Mail : Ten adres pocztowy jest chroniony przed spamowaniem. Aby go zobaczyć, konieczne jest włączenie w przeglądarce obsługi JavaScript. 

Strona głównaStrona głównaseparatorWspółpraca naukowa

Współpraca naukowa z zakresu fonoskopii

konferencja telefoniczna

Współpracuję z dr Bernardasem Šalna, ekspertem fonoskopii i byłym naczelnikiem fonoskopii Instytutu Ekspertyz Sądowych Litwy, który opracował metodykę badań pozwalającą zidentyfikować osoby posługujące się różnymi językami na podstawie głosu.

Dzięki tej metodyce możliwa jest identyfikacja osób różnej narodowości, np. działających w międzynarodowych grupach przestępczych.

 

 

First name and surname: Dr. Bernardas Šalna

 

Education and scientific degree:

1977 Graduated from Kaunas Polytechnic Institute (Kaunas Technology University – now), Dipl. in Radio Engineering;

 

1990 PhD in Development new methods for LPC synthesis of natural - sounding voiced speech for secure communication (in Russian language), Technical Cybernetics and Information Theory, Kaunas Polytechnic Institute, the Highest Attestation Commission, Moscow;

 

Academic title and employment:

1977-1991 Engineer Electronic, Chief Engineer Electronics at Vilnius University;

since 1991-2019 Head of the Department of Phonoscopic Expertise (Department of Speech and Audio Analysis ) Lithuanian Institute of Forensic Examination (actual Forensic Science Center of Lithuania ) and the lecturer of Criminalistic Department of Mykolas Riomeris University. 2020 - Senior Researcher of the Department of Speech and Audio Analysis.

 

The main fields of scientific interest:

 

Since 1977 working in the Speech analysis; Speech production and audio analysis; authentication of Forensic audio recordings, speaker recognition, voice biometrics and participate in multiple product

development projects for different applications. Bernardas Salna has been conducting experimental and scientific work of Phonoscopic Expertise Department and preparing the experts for independent work recently. Under the supervision of Bernardas Salna in the Forensic Science Center of Lithuania there has been created nine automated work places for the fulfillment of phonoscopic researches. The main result of his scientific work is creating methodological Guidelines for working with Forensic semiautomatic Speaker recognition package – SIVE. This package is used in work of experts from Finland, Estonia, Latvia, Poland, Czech, Romania, Greece, Portugal, Ukraine. Besides, on invitation of these countries he went to teach the experts of these countries. In 2004 on the invitation of Forensic Science Institute of Kiev (Ukraine ) upon the recommendation of ENFSI he took part in the work of international expert commission fulfilling the phonoscopic expertise in the murder case of journalist E. Gongadze. In 2017 on the invitation BELLINCAT perform forensic voice analysis in investigation into downing of Malaysian airlines flight 17.

 

Research publications:

Author of more than 40 papers, published in scientific journals and proceedings of conferences and patents “Method and System for Biometric –Voice Print Authentication “, US 7 536 304, US 8 280 740.

Active participation in “Lithuanian Language in Information Society 2000-2006” program and EU project " Real time network, text, and speaker analytics for combating organized crime (ROXANNE- www. roxanne-europroject. org).

Membership of professional associations:

Member of:

ELSNET - European Speech and Language Network),

IEEE - The Institute of Electrical and Electronics Engineers, Inc. ,

ENFSI -European Network of Forensic Science Institutes, Forensic Speech and Audio Analysis Work Group.

 

List of main scientific publications (enclosed).

 

Knowledge of foreign languages:

English (good, spoken and written); Russian(very good spoken and written ); German(good, spoken and written)

 

 

 

 

REPRESENTATIVE PUBLICATIONS

 

 

  1. www. roxanne-euproject. org. Forensic Automatic Speaker Recognition(FASR): Problems and prospects, 2020.

  2. www. roxanne-euproject. org. Overview of LTEC Databases & ASR system training, 2020.

  3. B. Šalna. Criminalistic person by voice identification system. Abstracts of second European Academy of forensic sciences Meeting. Poland, Krakow, 2000, p. 223.

  4. B. Šalna. Criminalistic person by voice identification system. Problems of Forensic Sciences. Second European Academy of Forensic Science Meeting, Cracow, September 12-16, 2000, Institute of Research Cracow, Poland, vol. XLVII p. 268-272.

  5. B. Šalna. Investigation of Reliability of Person Identification by Voice System SIVE, using Imitative Voice Base. 3 rd EUROPEAN ACADEMY OF FORENSIC SCIENCE TRIENNIAL MEETING. September 22 – 27, 2003, Istambul, Turkey, p. 364 – 365.

  6. B. Šalna. Validation of Forensic Speaker Identification System SIVE, using Twins Voice Base. Report of 7 –th Meeting of the ENFSI Expert WG for Forensic Speech and Audio Analysis (FSAAWG ), Toledo, Spain, May 28 – 29, 2004.

  7. B. Šalna. Use of Technologies of speaker recognition by voice in criminology and informatikon protection. The 9 – th International Conference of Innovative Technologies for Busines and Education. October 2007, Vilnius, Lithunia, p. 57-58.

9. B. Šalna. Voice biometrics – evaluation of effectiveness of different methods in Speaker recognition. May 16-17, 2008, Warsaw, Poland. The IEEE Workshop on Bio – Inspired Signal and Image Processing. BISIP 08. p. 48-51.

10. B. Šalna, J. Kamarauskas. Evaluation of Effectiveness of Different Methods in Speaker Recognition. Electronics and Electrical Engineering. 2010. No. 2(98), p. 67-70. ISSN. 1392-1215 Electronics and Electrical Engineering. Kaunas, Technologija.

  1. A. Bastys, A. Kisel, B. Šalna. The Use of Group Delay Features of Linear Prediction Model for Speaker Recognition. INFORMATICA, 2010, Vol. 21, No. 1, p. 1-12. ISSN 0868-4952, Institute of Mathematics and Informatics, Vilnius.

  2. B. Šalna, J. Kamarauskas. Speaker Recognition using Excitation Source Parameters. Electronics and Electrical Engineering, 2011, No. 1(107), p. 55-58, ISSN 1392- 1215 Kaunas, Lithuania.

  3. T. Tišina, E. Vaičiukynas, B. Šalna. Evaluation of effectiveness of Vector Quantization (VQ ) and 2D Discrete Cosine Transform(DCT2D) Methods for Speaker Authentification. Preceedings of the 9-th International Conference on Electrical and Control Technologies ECT-2014. Kaunas, Kaunas University of Technology, 2014, p. 36-41.

  4. B. Šalna, L. Lyvienė, A. Kowalczyk. Identification of person, talking in different languages, by the voice. XI INTERNATIONAL SCIENTIFIC – PRACTICAL CONFERENCE: Criminalistics and forensic science: science, studies, practice, 25-27 June, 2015, Vilnius, Lithuania, Mykolas Romeris University, p. 329-340 (In Lithuanian ).

  5. B. Šalna. Comparison of four FASR systems using real forensic voicedata. Report of 19 –th Meeting of the ENFSI Expert WG for Forensic Speech and Audio Analysis (FSAAWG ), Lisbon, Portugal, September 14 – 15, 2017.

  6. B. Šalna, J. Devenson, R. Konarskienė. Evaluation and Comparison of the Accuracy of two Forensic Automatic Speaker Recognition (FASR) system BATVOX and VOICE using real Forensic Voicedata. Proceedings of XVI INTERNATIONAL SCIENTIFIC – PRACTICAL CONFERENCE: Criminalistics and forensic science: science, studies, practice, 2020, Vilnius, Lithuania, Mykolas Romeris University, p. 232-250 (In Lithuanian).