Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Ensemble learning model for deepfake audio detection using multi-feature extraction approach.

Parna Chaudhury1, L Shivani Narayan1, S Mridhula1

  • 1School of Electronics Engineering (SENSE), Vellore Institute of Technology , Chennai, India.

Scientific Reports
|June 22, 2026
PubMed
Summary

Related Concept Videos

Extraction: Advanced Methods00:56

Extraction: Advanced Methods

Metal ions can be separated from one another by complexation with organic ligands–the chelating agent– to form uncharged chelates. Here, the chelating agent must contain hydrophobic groups and behave as a weak acid, losing a proton to bind with the metal. Since most organic ligands used in this process are insoluble or undergo oxidation in the aqueous phase, the chelating agent is initially added to the organic phase and extracted into the aqueous phase. The metal-ligand complex is formed in...

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Complete genome sequence of the carotenoid producing halophilic archaeon <i>Haloferax volcanii</i> TM<sup>T</sup>.

Microbiology resource announcements·2026
Same author

Damping factor estimation of damped complex sinusoidal signals using a maximum likelihood approach.

Scientific reports·2026
Same author

Dissecting mitogenomic conflict to illuminate angiosperm deep phylogeny: Sequence and architectural evidence.

Plant diversity·2026
Same author

An enhanced approach to minimum variance unbiased velocity estimation, incorporating horizontal and vertical handoff in HetNets.

Scientific reports·2025
Same author

Complete genome sequence of a <i>Salimicrobium</i> sp. PL1-032A isolated from the pink hypersaline Pearse Lakes, Rottnest Island, Western Australia.

Microbiology resource announcements·2024
Same author

Complete genome sequence of a <i>Marinococcus</i> sp. PL1-022 isolated from the pink hypersaline Pearse Lakes, Rottnest Island, Western Australia.

Microbiology resource announcements·2024

This study introduces an advanced ensemble learning method for improved deepfake detection. By fusing spectral, temporal, and statistical audio features, the system achieves significantly enhanced accuracy in identifying fake audio content.

Area of Science:

  • Artificial Intelligence
  • Machine Learning
  • Signal Processing

Background:

  • Deepfake technology poses a significant threat due to its potential for misinformation.
  • Accurate detection of audio deepfakes remains a challenge, requiring sophisticated analytical methods.

Purpose of the Study:

  • To enhance the accuracy and robustness of audio deepfake detection.
  • To explore the efficacy of ensemble learning by integrating diverse audio features.

Main Methods:

  • Leveraging Mel-frequency cepstral coefficients (MFCCs), statistical features, and short-time Fourier transform (STFT) features for audio analysis.
  • Employing specialized neural networks: MLPs for statistical features, RNNs for temporal MFCC analysis, and CNNs for STFT spectral patterns.
  • Implementing a stacking ensemble method with a meta-learning model to combine predictions from individual networks.
Keywords:
Audio forensicsDeepfake detectionEnsemble learningMFCCMeta-learningNeural networksSTFT

Related Experiment Videos

Main Results:

  • The proposed ensemble model demonstrated superior performance compared to individual models on the Fake-or-Real dataset.
  • Fusion of spectral, temporal, and statistical audio data significantly improved deepfake detection accuracy.
  • The system achieved highly accurate results in experimental evaluations.

Conclusions:

  • Ensemble learning integrating diverse audio features offers a robust approach to audio deepfake detection.
  • The combined analysis of spectral, temporal, and statistical characteristics is crucial for high-accuracy detection.
  • The developed model, accessible via Gradio on Hugging Face, provides real-time deepfake detection capabilities.