Publicación:
On the optimal selection of Mel-Frequency Cepstral Coefficients for voice deepfake detection

Vista sencilla de ítem
dc.contributor.authorFalcón López, Sergio A.
dc.contributor.authorTobarra Abad, María de los Llanos
dc.contributor.authorRobles Gómez, Antonio
dc.contributor.authorPastor Vargas, Rafael
dc.date.accessioned2026-03-09T11:00:57Z
dc.date.available2026-03-09T11:00:57Z
dc.date.issued2026-03-24
dc.descriptionThis is the Accepted Manuscript of an article published by in Expert Systems, Wiley; available online at the publisher's website: https://doi.org/10.1111/exsy.70245
dc.descriptionEste es el manuscrito aceptado de un artículo publicado en Expert Systems, Wiley; disponible en línea en el sitio web del editor: https://doi.org/10.1111/exsy.70245
dc.description.abstractThe continuous evolution of techniques for generating manipulated audio, known as voice deepfakes, and the widespread availability of tools that produce convincing forgeries have created an urgent need for reliable detection methods. This work considers the dimensionality of Mel-Frequency Cepstral Coefficients (MFCCs) as a core design variable for practical, deployable systems. The aim is to identify the smallest number of coefficients that preserves detection performance across heterogeneous models while reducing computational cost, a critical factor for mobile and edge deployment. This study evaluates a hybrid setting on the ASVspoof 2019 Logical Access dataset, in which the same feature family serves as input to five traditional machine learning algorithms (Random Forest, k-Nearest Neighbors, Linear Support Vector Classification, Extreme Gradient Boosting and Support Vector Machine with radial basis function kernel) and five deep learning models (Convolutional Neural Network, Recurrent Neural Network, Convolutional Recurrent Neural Network, Xception and ResNet). Results indicate that deep models reach near-peak performance with a small number of coefficients, whereas classical methods require a larger number to achieve stable performance (except Linear Support Vector Classification, which consistently underperforms). Accordingly, 32 coefficients are considered an effective operating point for hybrid deployments. Overall, the results provide evidence to guide the selection of the number of MFCC coefficients in voice deepfake detection, aiming for efficient, reproducible and explainable systems.en
dc.description.provenanceMade available in DSpace on 2026-03-09T11:00:57Z (GMT). No. of bitstreams: 1 RoblesGomez_Antonio_On-the-optimal-selection-_ANTONIO ROBLES GOMEZ.pdf: 873508 bytes, checksum: 83c732856db07cc330dee060ba62aa14 (MD5) Previous issue date: 2026-03-06en
dc.description.versionversión final
dc.identifier.citationFalcón-López, S.A., Tobarra, L., Robles- Gómez, A., Pastor-Vargas, R. (2026); On the optimal selection of Mel-Frequency Cepstral Coefficients for voice deepfake detection; Publicación: Expert Systems; Wiley, ; Páginas 1-33, https://doi.org/10.1111/exsy.70245
dc.identifier.doihttps://doi.org/10.1111/exsy.70245
dc.identifier.eissn1468-0394
dc.identifier.issn0266-4720
dc.identifier.urihttps://hdl.handle.net/20.500.14468/32049
dc.journal.titleExpert Systems
dc.language.isoen
dc.publisherWiley
dc.relation.centerEscuela Técnica Superior de Ingeniería Informática
dc.relation.departmentSistemas de Comunicación y Control
dc.rightsinfo:eu-repo/semantics/openAccess
dc.rights.urihttp://creativecommons.org/licenses/by-nc-nd/4.0/deed.es
dc.subject1203.18 Sistemas de información, diseño y componentes
dc.subject.keywordsDeepfakeen
dc.subject.keywordsForensic Analysisen
dc.subject.keywordsAudio deepfake detectionen
dc.subject.odsODS 9 - Industria, innovación e infraestructura
dc.titleOn the optimal selection of Mel-Frequency Cepstral Coefficients for voice deepfake detectionen
dc.typeartículoes
dc.typejournal articleen
dspace.entity.typePublication
relation.isAuthorOfPublicationb584f8a3-eb01-4a43-9ed7-5075b74224ae
relation.isAuthorOfPublication17556659-f434-4220-841d-aac35f492e62
relation.isAuthorOfPublicationf93103de-336d-47ac-886b-e2cbd425ed87
relation.isAuthorOfPublication.latestForDiscoveryb584f8a3-eb01-4a43-9ed7-5075b74224ae
Archivos
Bloque original
Mostrando 1 - 1 de 1
Miniatura
Nombre:
RoblesGomez_Antonio_On-the-optimal-selection-_ANTONIO ROBLES GOMEZ.pdf
Tamaño:
853.04 KB
Formato:
Adobe Portable Document Format
Descargar
Bloque de licencias
Mostrando 1 - 1 de 1
No hay miniatura disponible
Nombre:
license.txt
Tamaño:
3.62 KB
Formato:
Item-specific license agreed to upon submission
Descripción:
Descargar
Colecciones
Artículos y papers