Objective Control for TAlker VErification

Livrables

Suitability of existing voice biometric engines

This deliverable outlines the current state of the main areas of research relevant to design phase and development. It consists of an extended technical report that reviews in particular the leading-edge state-of-the-art voice biometric engines (both commercial and non-commercial), while considering their pros and cons from the viewpoint of OCTAVE project goals. These include robustness to background noise and channel variation, resistance against spoofing attacks, and recognition accuracy on short utterances.

OCTAVE Report on communication actions and participation in events Y2

This report provides a list of journals, publications, bulletins, conferences, events and workshops where OCTAVE results will have been communicated during the second year of the project. It also includes a description of the dissemination material created to promote the project in the different attended events.

Changes to platform for Privacy compliance

This deliverable demonstrates how biometric data can be anonymized in an authentication environment. The sample platform uses multifactor authentication, where a speaker verification engine (SV) provides a strong identity factor when required. A well-supported federation protocol is used to insulate the SV both physically and logically from other identity-related data, like emails and bank accounts. The SV communicates with the authentication platform by using a token proxy, which represents the end user's identity. The ultimate goal is to reduce exposure to a single point of failure, if data leakages do happen.

Baseline voice biometric systems and their vulnerability to spoofing

This is the publishable version of D3. This deliverable is a report setting out the baseline voice biometric technologies considered for the project and analyzing their vulnerabilities to spoofing. The types of spoofing attacks presented in the report are primarily in the context of the operational modes of speaker verification which are adopted for the considered application areas.

OCTAVE report on communication actions and participation in events Y1

This report provides a list of journals, publications, bulletins, conferences, events and workshops where OCTAVE results have been communicated during the first year of the project. It also includes a description of the dissemination material created to promote the project in the different attended events.

Contribution to standards

This report will be an organised and commented collection of all contributions that the Project will have made to standard bodies, at national, European and global level. The contributions themselves will be part of the Deliverable.

Survey of the safety and security standards, procedures and technologies for the collection, transmission, storage and processing of personal data

This document is a survey of the safety and security standards, procedures and technologies for the collection, transmission, storage and processing of personal data. Contributing to WP2 as well as to the general privacy-by-design approach in the project, this document will be the key for OCTAVE to adopt standard, secure formats for the exchange of biometric data and authentication requests/responses, such as the recently defined OpenID Connect standard.

Corpora definition

This Deliverable describes the OCTAVE corpora definition and collection method on the basis of preliminary requirements set for trial design. It will define the various corpora needed and, if necessary, an extra integrative corpus to be implemented. Among the requirements we will also include the recommendations issued from the Ethics Committee for collection of ad hoc corpora. The expected result is a database of appropriate size, able to provide the best representation of data and boundary conditions typically encountered in real scenarios involving variations of spoken language and communication channels.

OCTAVE Website

This deliverable is twofold: a report on the design of the OCTAVE website and the initial website itself (home page and some basic pages, as of indicated delivery month). The website is of course a living product, as it will be constantly informing the target audiences about the existence of the OCTAVE project and the potential benefits for companies and experts operating in the expected fields. The website will make available social-network targeted information on the themes of access control and biometric data. The website contents will address the general public (by presenting the advantages of voice biometrics for easy access to information services and critical infrastructures) and all stakeholders (e.g., enterprises requiring access control, smart service developers, educators, researchers, etc.) by providing them with outputs that are publishable according to OCTAVE classification policies.

Publications

Local spectral variability features for speaker verification

Auteurs: Md Sahidullah, Tomi Kinnunen
Publié dans: Digital Signal Processing, Numéro 50, 2016, Page(s) 1-11, ISSN 1051-2004
Éditeur: Academic Press
DOI: 10.1016/j.dsp.2015.10.011

Spoofing detection goes noisy: An analysis of synthetic speech detection in the presence of additive noise

Auteurs: Cemal Hanilçi, Tomi Kinnunen, Md Sahidullah, Aleksandr Sizov
Publié dans: Speech Communication, Numéro 85, 2016, Page(s) 83-97, ISSN 0167-6393
Éditeur: Elsevier BV
DOI: 10.1016/j.specom.2016.10.002

ASVspoof: The Automatic Speaker Verification Spoofing and Countermeasures Challenge

Auteurs: Zhizheng Wu, Junichi Yamagishi, Tomi Kinnunen, Cemal Hanilci, Mohammed Sahidullah, Aleksandr Sizov, Nicholas Evans, Massimiliano Todisco
Publié dans: IEEE Journal of Selected Topics in Signal Processing, Numéro 11/4, 2017, Page(s) 588-604, ISSN 1932-4553
Éditeur: Institute of Electrical and Electronics Engineers
DOI: 10.1109/JSTSP.2017.2671435

Semi-supervised speech activity detection with an application to automatic speaker verification

Auteurs: Alexey Sholokhov, Md Sahidullah, Tomi Kinnunen
Publié dans: Computer Speech & Language, Numéro 47, 2018, Page(s) 132-156, ISSN 0885-2308
Éditeur: Academic Press
DOI: 10.1016/j.csl.2017.07.005

DNN Filter Bank Cepstral Coefficients for Spoofing Detection

Auteurs: Hong Yu, Zheng-Hua Tan, Yiming Zhang, Zhanyu Ma, Jun Guo
Publié dans: IEEE Access, Numéro 5, 2017, Page(s) 4779-4787, ISSN 2169-3536
Éditeur: Institute of Electrical and Electronics Engineers Inc.
DOI: 10.1109/ACCESS.2017.2687041

Incorporating pass-phrase dependent background models for text-dependent speaker verification

Auteurs: A.K. Sarkar, Zheng-Hua Tan
Publié dans: Computer Speech & Language, 2017, ISSN 0885-2308
Éditeur: Academic Press
DOI: 10.1016/j.csl.2017.07.010

Constant Q cepstral coefficients: A spoofing countermeasure for automatic speaker verification

Auteurs: Massimiliano Todisco, Héctor Delgado, Nicholas Evans
Publié dans: Computer Speech & Language, Numéro 45, 2017, Page(s) 516-535, ISSN 0885-2308
Éditeur: Academic Press
DOI: 10.1016/j.csl.2017.01.001

Robust Voice Liveness Detection and Speaker Verification Using Throat Microphones

Auteurs: Md Sahidullah, Dennis Alexander Lehmann Thomsen, Rosa Gonzalez Hautamaki, Tomi Kinnunen, Zheng-Hua Tan, Robert Parts, Martti Pitkanen
Publié dans: IEEE/ACM Transactions on Audio, Speech, and Language Processing, 2017, Page(s) 1-1, ISSN 2329-9290
Éditeur: IEEE Advancing Technology for Humanity
DOI: 10.1109/taslp.2017.2760243

A New Feature for Automatic Speaker Verification Anti-Spoofing: Constant Q Cepstral Coefficients

Auteurs: Massimiliano Todisco, Héctor Delgado, Nicholas Evans
Publié dans: Odyssey 2016, 2016, Page(s) 283-290
Éditeur: ISCA
DOI: 10.21437/Odyssey.2016-41

Integrated Spoofing Countermeasures and Automatic Speaker Verification: An Evaluation on ASVspoof 2015

Auteurs: Md. Sahidullah, Héctor Delgado, Massimiliano Todisco, Hong Yu, Tomi Kinnunen, Nicholas Evans, Zheng-Hua Tan
Publié dans: Interspeech 2016, 2016, Page(s) 1700-1704
Éditeur: ISCA
DOI: 10.21437/Interspeech.2016-1280

Performance comparison of speaker recognition systems in presence of duration variability

Auteurs: Arnab Poddar, Md Sahidullah, Goutam Saha
Publié dans: 2015 Annual IEEE India Conference (INDICON), 2015, Page(s) 1-6, ISBN 978-1-4673-7399-9
Éditeur: IEEE
DOI: 10.1109/INDICON.2015.7443464

Classifiers for synthetic speech detection: a comparison

Auteurs: C. Hanilçi, T. Kinnunen, M. Sahidullah, A. Sizov
Publié dans: Proc. Interspeech 2015, 2015
Éditeur: ISCA

ASVspoof 2015: the first automatic speaker verification spoofing and countermeasures challenge

Auteurs: Z. Wu, T. Kinnunen, N. Evans, J. Yamagishi, C. Hanilçi, M. Sahidullah, A. Sizov
Publié dans: Proc. Interspeech 2015, 2015
Éditeur: ISCA

Speaker recognition for speech under face cover

Auteurs: R. Saeidi, T. Niemi, H. Karppelin, J. Pohjalainen, T. Kinnunen, P. Alku
Publié dans: Proc. Interspeech 2015, 2015
Éditeur: ISCA

Discriminative multi-domain PLDA for speaker verification

Auteurs: Alexey Sholokhov, Tomi Kinnunen, Sandro Cumani
Publié dans: 2016 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2016, Page(s) 5030-5034, ISBN 978-1-4799-9988-0
Éditeur: IEEE
DOI: 10.1109/ICASSP.2016.7472635

Age-Related Voice Disguise and its Impact on Speaker Verification Accuracy

Auteurs: Rosa González Hautamäki, Md Sahidullah, Tomi Kinnunen, Ville Hautamäki
Publié dans: Odyssey 2016, 2016, Page(s) 277-282
Éditeur: ISCA
DOI: 10.21437/Odyssey.2016-40

A comparison of features for synthetic speech detection

Auteurs: M. Sahidullah, T. Kinnunen, C. Hanilçi
Publié dans: Proc. Interspeech 2015, 2015
Éditeur: ISCA

Utterance Verification for Text-Dependent Speaker Recognition: A Comparative Assessment Using the RedDots Corpus

Auteurs: Tomi Kinnunen, Md. Sahidullah, Ivan Kukanov, Héctor Delgado, Massimiliano Todisco, Achintya Kr. Sarkar, Nicolai Bæk Thomsen, Ville Hautamäki, Nicholas Evans, Zheng-Hua Tan
Publié dans: Interspeech 2016, 2016, Page(s) 430-434
Éditeur: ISCA
DOI: 10.21437/Interspeech.2016-1125

Articulation Rate Filtering of CQCC Features for Automatic Speaker Verification

Auteurs: Massimiliano Todisco, Héctor Delgado, Nicholas Evans
Publié dans: Interspeech 2016, 2016, Page(s) 3628-3632
Éditeur: ISCA
DOI: 10.21437/Interspeech.2016-1140

Robust Speaker Recognition with Combined Use of Acoustic and Throat Microphone Speech

Auteurs: Md. Sahidullah, Rosa Gonzalez Hautamäki, Dennis Alexander Lehmann Thomsen, Tomi Kinnunen, Zheng-Hua Tan, Ville Hautamäki, Robert Parts, Martti Pitkänen
Publié dans: Interspeech 2016, 2016, Page(s) 1720-1724
Éditeur: ISCA
DOI: 10.21437/Interspeech.2016-1153

Speaker-Dependent Dictionary-Based Speech Enhancement for Text-Dependent Speaker Verification

Auteurs: Nicolai Bæk Thomsen, Dennis Alexander Lehmann Thomsen, Zheng-Hua Tan, Børge Lindberg, Søren Holdt Jensen
Publié dans: Interspeech 2016, 2016, Page(s) 1839-1843
Éditeur: ISCA
DOI: 10.21437/Interspeech.2016-763

Speech enhancement using Long Short-Term Memory based recurrent Neural Networks for noise robust Speaker Verification

Auteurs: Morten Kolboek, Zheng-Hua Tan, Jesper Jensen
Publié dans: 2016 IEEE Spoken Language Technology Workshop (SLT), 2016, Page(s) 305-311, ISBN 978-1-5090-4903-5
Éditeur: IEEE
DOI: 10.1109/SLT.2016.7846281

Further optimisations of constant Q cepstral processing for integrated utterance and text-dependent speaker verification

Auteurs: Hector Delgado, Massimiliano Todisco, Md Sahidullah, Achintya K. Sarkar, Nicholas Evans, Tomi Kinnunen, Zheng-Hua Tan
Publié dans: 2016 IEEE Spoken Language Technology Workshop (SLT), 2016, Page(s) 179-185, ISBN 978-1-5090-4903-5
Éditeur: IEEE
DOI: 10.1109/SLT.2016.7846262

Effects of gender information in text-independent and text-dependent speaker verification

Auteurs: Anssi Kanervisto, Ville Vestman, Md Sahidullah, Ville Hautamaki, Tomi Kinnunen
Publié dans: 2017 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2017, Page(s) 5360-5364, ISBN 978-1-5090-4117-6
Éditeur: IEEE
DOI: 10.1109/ICASSP.2017.7953180

Generalization of spoofing countermeasures: A case study with ASVspoof 2015 and BTAS 2016 corpora

Auteurs: Dipjyoti Paul, Md Sahidullah, Goutam Saha
Publié dans: 2017 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2017, Page(s) 2047-2051, ISBN 978-1-5090-4117-6
Éditeur: IEEE
DOI: 10.1109/ICASSP.2017.7952516

The I4U Mega Fusion and Collaboration for NIST Speaker Recognition Evaluation 2016

Auteurs: Kong Aik Lee, SRE Group
Publié dans: Interspeech 2017, 2017, Page(s) 1328-1332
Éditeur: ISCA
DOI: 10.21437/Interspeech.2017-203

The ASVspoof 2017 Challenge: Assessing the Limits of Replay Spoofing Attack Detection

Auteurs: Tomi Kinnunen, Md. Sahidullah, Héctor Delgado, Massimiliano Todisco, Nicholas Evans, Junichi Yamagishi, Kong Aik Lee
Publié dans: Interspeech 2017, 2017, Page(s) 2-6
Éditeur: ISCA
DOI: 10.21437/Interspeech.2017-1111

Improving Speaker Verification Performance in Presence of Spoofing Attacks Using Out-of-Domain Spoofed Data

Auteurs: Achintya Kr. Sarkar, Md. Sahidullah, Zheng-Hua Tan, Tomi Kinnunen
Publié dans: Interspeech 2017, 2017, Page(s) 2611-2615
Éditeur: ISCA
DOI: 10.21437/Interspeech.2017-1758

Time-Varying Autoregressions for Speaker Verification in Reverberant Conditions

Auteurs: Ville Vestman, Dhananjaya Gowda, Md. Sahidullah, Paavo Alku, Tomi Kinnunen
Publié dans: Interspeech 2017, 2017, Page(s) 1512-1516
Éditeur: ISCA
DOI: 10.21437/Interspeech.2017-734

Adversarial Network Bottleneck Features for Noise Robust Speaker Verification

Auteurs: Hong Yu, Zheng-Hua Tan, Zhanyu Ma, Jun Guo
Publié dans: Interspeech 2017, 2017, Page(s) 1492-1496
Éditeur: ISCA
DOI: 10.21437/Interspeech.2017-883

Conditional Generative Adversarial Networks for Speech Enhancement and Noise-Robust Speaker Verification

Auteurs: Daniel Michelsanti, Zheng-Hua Tan
Publié dans: Interspeech 2017, 2017, Page(s) 2008-2012
Éditeur: ISCA
DOI: 10.21437/Interspeech.2017-1620

Improving Speaker Verification Performance under Spoofing Attacks By Fusion of Different Operational Modes

Auteurs: S. Safavi, I. Mporas, and H. Gan
Publié dans: Conference IEEE CSPA, 2017
Éditeur: IEEE

Fraud Detection in Voice-Based Identity Authentication Applications and Services

Auteurs: Saeid Safavi, Hock Gan, Iosif Mporas, Reza Sotudeh
Publié dans: 2016 IEEE 16th International Conference on Data Mining Workshops (ICDMW), 2016, Page(s) 1074-1081, ISBN 978-1-5090-5910-2
Éditeur: IEEE
DOI: 10.1109/ICDMW.2016.0155

Combination or Rule-Based and Data-Driven Fusion Methodologies for Different Speaker Verification Modes of Operation

Auteurs: S. Safavi and I. Mporas
Publié dans: Conference: IEEE CPSCom, 2017
Éditeur: IEEE

Effect of multi-condition training and speech enhancement methods on spoofing detection

Auteurs: Hong Yu, Achintya Sarkar, Dennis Alexander Lehmann Thomsen, Zheng-Hua Tan, Zhanyu Ma, Jun Guo
Publié dans: 2016 First International Workshop on Sensing, Processing and Learning for Intelligent Machines (SPLINE), 2016, Page(s) 1-5, ISBN 978-1-4673-8917-4
Éditeur: IEEE
DOI: 10.1109/SPLIM.2016.7528399

HAPPY Team Entry to NIST OpenSAD Challenge: A Fusion of Short-Term Unsupervised and Segment i-Vector Based Speech Activity Detectors

Auteurs: Tomi Kinnunen, Alexey Sholokhov, Elie Khoury, Dennis Alexander Lehmann Thomsen, Md. Sahidullah, Zheng-Hua Tan
Publié dans: Interspeech 2016, 2016, Page(s) 2992-2996
Éditeur: ISCA
DOI: 10.21437/Interspeech.2016-1281

Evaluation of Classification Algorithms for Text Dependent and Text Independent Speaker Identification

Auteurs: Iosif Mporas , Saeid Safavi , Hock Chye Gan, Reza Sotudeh
Publié dans: Proceedings of 2016 IEICE ICTF, 2016
Éditeur: i-scover

RedDots replayed: A new replay spoofing attack corpus for text-dependent speaker verification research

Auteurs: Tomi Kinnunen, Md Sahidullah, Mauro Falcone, Luca Costantini, Rosa Gonzalez Hautamaki, Dennis Thomsen, Achintya Sarkar, Zheng-Hua Tan, Hector Delgado, Massimiliano Todisco, Nicholas Evans, Ville Hautamaki, Kong Aik Lee
Publié dans: 2017 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2017, Page(s) 5395-5399, ISBN 978-1-5090-4117-6
Éditeur: IEEE
DOI: 10.1109/ICASSP.2017.7953187

Text Dependent Speaker Verification Using Un-Supervised HMM-UBM and Temporal GMM-UBM

Auteurs: Achintya Kr. Sarkar, Zheng-Hua Tan
Publié dans: Interspeech 2016, 2016, Page(s) 425-429
Éditeur: ISCA
DOI: 10.21437/Interspeech.2016-362

ASVspoof 2015: The First Automatic Speaker Verification Spoofing and Countermeasures Challenge

Auteurs: Zhizheng Wu, Tomi Kinnunen, Nicholas Evans, and Junichi Yamagishi
Publié dans: SLTC Newsletter, November 2015, 2015
Éditeur: IEEE

Improving Performance of Speaker Identification Systems Using Score Level Fusion of Two Modes of Operation

Auteurs: Saeid Safavi, Iosif Mporas
Publié dans: SPECOM 2017. Lecture Notes in Computer Science, vol 10458, 2016, Page(s) 438-444
Éditeur: Springer International Publishing
DOI: 10.1007/978-3-319-66429-3_43

Improving Robustness of Speaker Verification by Fusion of Prompted Text-Dependent and Text-Independent Operation Modalities

Auteurs: Iosif Mporas, Saeid Safavi, Reza Sotudeh
Publié dans: Book, 2015, Page(s) 378-385, ISBN 978-3-319-43957-0
Éditeur: Springer International Publishing
DOI: 10.1007/978-3-319-43958-7_45

Droits de propriété intellectuelle

Audiovisual associative authentication method, related system and device

Numéro de demande/publication: BP 209598PC
Date: 2015-03-20
Demandeur(s): APLCOMP OY

Audiovisual associative authentication method and related system.

Numéro de demande/publication: FI 20155197
Date: 2015-03-20
Demandeur(s): APLCOMP OY

Audiovisual associative authentication method and related system.

Numéro de demande/publication: FI 20155197
Date: 2015-03-20
Demandeur(s): APLCOMP OY

"Kaksoismikrofonilaite ja tämän sisätävä kuulokemikrofoni (""Dual microphone with integrated earphone"")"

Numéro de demande/publication: FI 11241
Date: 2015-12-17
Demandeur(s): APLCOMP OY

Livrables

Publications

Droits de propriété intellectuelle

Partager cette page

Télécharger