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

Sensors-Driven Multimodal Deepfake Detection: A Cross-Attention Fusion Approach with Adaptive Modality Gating.

Syeda Sitara Waseem1, Noman Shabbir2, Syed Rizwan Hassan3

  • 1Department of Computer Science & IT, The Government Sadiq College Women University, Bahawalpur 63100, Pakistan.

Sensors (Basel, Switzerland)
|June 26, 2026
PubMed
Summary

Related Concept Videos

Masking and Demasking Agents01:19

Masking and Demasking Agents

EDTA titrations may necessitate masking and demasking agents to temporarily protect a particular metal ion in a mixture from the EDTA reaction. These agents facilitate the sequential analysis of the metal ions by forming stable complexes with some—but not all—metal ions during certain steps.
There are many masking agents, such as cyanide, fluoride, triethanolamine, thiourea, and 2,3-bis(sulfanyl)propan-1-ol (formerly 2,3-dimercapto-1-propanol), with the masking agent chosen based on the metal...

You might also read

Related Articles

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

Sort by
Same author

IMAU-Net: A Hybrid Multi-Scale Deep Learning Framework for Liver Segmentation from Laparoscopic Images.

Sensors (Basel, Switzerland)·2026
Same author

IoT-Oriented Security for Small Sensor Systems Using DnCNN Denoising and Multimodal Feature Fusion for Image Forgery Detection.

Sensors (Basel, Switzerland)·2026
Same author

QSEER-Quantum-Enhanced Secure and Energy-Efficient Routing Protocol for Wireless Sensor Networks (WSNs).

Sensors (Basel, Switzerland)·2025
Same author

Correction: Comprehensive Symptom Prediction in Inpatients With Acute Psychiatric Disorders Using Wearable-Based Deep Learning Models: Development and Validation Study.

Journal of medical Internet research·2024
Same author

Comprehensive Symptom Prediction in Inpatients With Acute Psychiatric Disorders Using Wearable-Based Deep Learning Models: Development and Validation Study.

Journal of medical Internet research·2024
Same author

A Review on Electronic Health Record Text-Mining for Biomedical Name Entity Recognition in Healthcare Domain.

Healthcare (Basel, Switzerland)·2023

This study introduces a multimodal deepfake detection framework for edge devices, achieving high accuracy against audio and video fakes. The efficient model offers robust security for sensor-based authentication systems.

Area of Science:

  • Computer Science
  • Artificial Intelligence
  • Cybersecurity

Background:

  • Deepfakes pose a significant threat to sensor-based authentication systems like cameras and IoT devices.
  • Existing unimodal detectors are insufficient against sophisticated, modality-specific deepfake attacks.

Purpose of the Study:

  • To develop an efficient multimodal deepfake detection framework for resource-constrained edge devices.
  • To enhance the security of sensor-based authentication against deepfake manipulations.

Main Methods:

  • Proposed a novel cross-modal attention fusion mechanism with adaptive gating for multimodal deepfake detection.
  • Integrated enhanced Res2Net for audio analysis and temporal 3D CNN with SE attention for video processing.
  • Employed bidirectional cross-modal attention with quality-based gates for robust fusion.
Keywords:
3D CNNIoT securityRes2Netadversarial robustnessaudiovisual sensorscross-modal attentiondeepfake detectionedge computingmultimodal sensor fusionreal-time biometric authenticationsensor-based forensicssmart camera systems

Related Experiment Videos

Main Results:

  • Achieved 96.7% accuracy, 96.6% F1-score, and 0.988 AUC-ROC on a benchmark dataset.
  • Demonstrated 92.3% accuracy against Fast Gradient Sign Method (FGSM) adversarial attacks.
  • Model exhibits a 30.3 MB footprint and operates at 20 FPS on edge hardware, with adaptive modality weighting.

Conclusions:

  • The developed multimodal framework offers efficient and accurate deepfake detection on edge devices.
  • The system shows strong generalization capabilities across different datasets and robustness against adversarial attacks.
  • Adaptive modality contribution highlights the model's flexibility in handling different types of deepfakes (e.g., TTS, lip-synced).