Toward Fail-Safe Speaker Recognition: Trial-Based Calibration with a Reject Option | ||
| dc.conicet.organizacion | Consejo Nacional de Investigaciones Científicas y Técnicas. Oficina de Coordinación Administrativa Ciudad Universitaria. Instituto de Investigación en Ciencias de la Computación. Universidad de Buenos Aires. Facultad de Ciencias Exactas y Naturales. Instituto de Investigación en Ciencias de la Computación | |
| dc.conicet.organizacion | No especifíca | |
| dc.conicet.organizacion | No especifíca | |
| dc.conicet.organizacion | No especifíca | |
| dc.conicet.organizacion | No especifíca | |
| dc.contributor.author | Ferrer, Luciana | |
| dc.contributor.author | Nandwana, Mahesh Kumar | |
| dc.contributor.author | McLaren, Mitchell | |
| dc.contributor.author | Castan, Diego | |
| dc.contributor.author | Lawson, Aaron | |
| dc.date.issued | 2019-01 | |
| dc.description.abstract | The output scores of most of the speaker recognition systems are not directly interpretable as stand-alone values. For this reason, a calibration step is usually performed on the scores to convert them into proper likelihood ratios, which have a clear probabilistic interpretation. The standard calibration approach transforms the system scores using a linear function trained using data selected to closely match the evaluation conditions. This selection, though, is not feasible when the evaluation conditions are unknown. In previous work, we proposed a calibration approach for this scenario called trial-based calibration (TBC). TBC trains a separate calibration model for each test trial using data that is dynamically selected from a candidate training set to match the conditions of the trial. In this work, we extend the TBC method, proposing: 1) a new similarity metric for selecting training data that result in significant gains over the one proposed in the original work; 2) a new option that enables the system to reject a trial when not enough matched data are available for training the calibration model; and 3) the use of regularization to improve the robustness of the calibration models trained for each trial. We test the proposed algorithms on a development set composed of several conditions and on the Federal Bureau of Investigation multi-condition speaker recognition dataset, and we demonstrate that the proposed approach reduces calibration loss to values close to 0 for most of the conditions when matched calibration data are available for selection, and that it can reject most of the trials for which relevant calibration data are unavailable. | es |
| dc.description.fil | Fil: Ferrer, Luciana. Consejo Nacional de Investigaciones Científicas y Técnicas. Oficina de Coordinación Administrativa Ciudad Universitaria. Instituto de Investigación en Ciencias de la Computación. Universidad de Buenos Aires. Facultad de Ciencias Exactas y Naturales. Instituto de Investigación en Ciencias de la Computación; Argentina | |
| dc.description.fil | Fil: Nandwana, Mahesh Kumar. No especifíca; | |
| dc.description.fil | Fil: McLaren, Mitchell. No especifíca; | |
| dc.description.fil | Fil: Castan, Diego. No especifíca; | |
| dc.description.fil | Fil: Lawson, Aaron. No especifíca; | |
| dc.format | application/pdf | es |
| dc.identifier.citation | Ferrer, Luciana; Nandwana, Mahesh Kumar; McLaren, Mitchell; Castan, Diego; Lawson, Aaron; Toward Fail-Safe Speaker Recognition: Trial-Based Calibration with a Reject Option; Institute of Electrical and Electronics Engineers; IEEE/ACM Transactions on Audio Speech and Language Processing; 27; 1; 1-2019; 140-153 | es |
| dc.identifier.institucional | CONICET | |
| dc.identifier.issn | 2329-9290 | es |
| dc.identifier.repo | CONICET Digital | |
| dc.identifier.uri | http://hdl.handle.net/11336/123318 | |
| dc.journal.number | 1 | es |
| dc.journal.pagination | 140-153 | es |
| dc.journal.pais | Estados Unidos | es |
| dc.journal.title | IEEE/ACM Transactions on Audio Speech and Language Processing | es |
| dc.journal.volume | 27 | es |
| dc.language.iso | eng | es |
| dc.publisher | Institute of Electrical and Electronics Engineers | es |
| dc.relation.alternativeid | info:eu-repo/semantics/altIdentifier/url/https://ieeexplore.ieee.org/document/8490592 | |
| dc.relation.alternativeid | info:eu-repo/semantics/altIdentifier/doi/http://dx.doi.org/10.1109/TASLP.2018.2875794 | |
| dc.rights | info:eu-repo/semantics/openAccess | * |
| dc.rights.uri | https://creativecommons.org/licenses/by-nc-sa/2.5/ar/ | * |
| dc.subject | FORENSIC VOICE COMPARISON | es |
| dc.subject | SPEAKER RECOGNITION | es |
| dc.subject | TRIAL-BASED CALIBRATION | es |
| dc.subject.classification | Ciencias de la Información y Bioinformática | es |
| dc.subject.classification | Ciencias de la Computación e Información | es |
| dc.subject.classification | CIENCIAS NATURALES Y EXACTAS | es |
| dc.title | Toward Fail-Safe Speaker Recognition: Trial-Based Calibration with a Reject Option | es |
| dc.type | info:eu-repo/semantics/article | |
| dc.type | info:ar-repo/semantics/artículo | |
| dc.type | info:eu-repo/semantics/publishedVersion | |
Archivos
Bloque original
Cargando...
- Nombre:
- CONICET_Digital_Nro.dedefff2-438f-4650-8ce7-d796f03b7764_A.pdf
- Tamaño:
- 2.26 MB
- Formato:
- Adobe Portable Document Format