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

DeepForgeryNet: a hybrid CNN-LSTM and transfer learning framework for robust image forgery and deepfake detection.

Aarti Sardhara1, Vipul Vekariya1, Ajeet Ram Pathak2

  • 1Department of Computer Science and Engineering, Faculty of Engineering and Technology, Parul Institute of Engineering and Technology, Parul University, Waghodia, Gujarat, India.

Frontiers in Artificial Intelligence
|May 25, 2026
PubMed
Summary

Related Concept Videos

You might also read

Related Articles

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

Sort by
Same author

Hybrid DenseNet-U-Net framework for automated grading of renal cell carcinoma.

Digital health·2026
Same author

Artificial intelligence for early endometrial cancer diagnosis using multimodal clinical data: integrating deep learning, explainability, and data privacy.

Frontiers in artificial intelligence·2026
Same author

Explainable multi-modal deep learning for transparent cancer diagnosis: integrating radiology, clinical features, and decision visualization.

Frontiers in artificial intelligence·2026
Same author

Intelligent Congestion Control Mechanism for IoT-Enabled Wireless Sensor Networks Using Hybrid Aggregation and Scheduling Technique.

Journal of visualized experiments : JoVE·2026
Same author

Explainable multilingual and multimodal fake-news detection: toward robust and trustworthy AI for combating misinformation.

Frontiers in artificial intelligence·2025
Same author

Environmental sustainability and waste conversion of <i>Prosopis juliflora</i> fibre-reinforced ZnO nanofiller particulates PLA composite- mechanical and thermal analysis.

Heliyon·2024
Same journal

Scale, trust, and the digital divide: a systematic review of AI and ML for agricultural applications.

Frontiers in artificial intelligence·2026
Same journal

Beyond uncertainty in modern active learning for trustworthy AI.

Frontiers in artificial intelligence·2026
Same journal

Eco-FinOps: a causal-agentic framework for energy-efficient and explainable cloud cost optimization.

Frontiers in artificial intelligence·2026
Same journal

Multimodal graph neural network with large language models for node and link prediction.

Frontiers in artificial intelligence·2026
Same journal

Efficient representation of boolean decision structures through Boolean function optimization.

Frontiers in artificial intelligence·2026
Same journal

Structural impact of non-IID heterogeneity on federated behavioral anomaly detection in IoT and IoMT systems.

Frontiers in artificial intelligence·2026
See all related articles
This summary is machine-generated.

A new deep learning model, DeepForgeryNet, effectively detects AI-generated and manipulated images by analyzing artifact and contextual clues. This advanced digital media verification tool achieves high accuracy, enhancing trust in online content.

Area of Science:

  • Computer Science
  • Artificial Intelligence
  • Digital Forensics

Background:

  • The proliferation of AI-generated and digitally manipulated images challenges media authenticity and digital trust.
  • Existing detection methods often overlook subtle forensic traces by focusing solely on visual content.

Purpose of the Study:

  • To develop a robust detection tool for identifying sophisticated image forgeries.
  • To leverage both artifact-level and contextual inconsistencies for enhanced detection accuracy.

Main Methods:

  • Introduction of DeepForgeryNet, an artifact-aware deep learning model.
  • Integration of Error Level Analysis (ELA) based preprocessing with a hybrid Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM) network.
  • End-to-end training on publicly available benchmark datasets with stratified data splits.
Keywords:
cross-dataset generalizationdeepfake analysisdigital image forensicserror level analysishybrid CNN–LSTMimage forgery detectiontransfer learning

Related Experiment Videos

Main Results:

  • Achieved high performance metrics: 95.1% accuracy, 94.6% precision, 94.2% recall, 94.4% F1-score, and 0.98 AUC.
  • Outperformed baseline CNN and transformer models, particularly in recall.
  • Demonstrated stable generalization with over 92% accuracy in cross-dataset experiments.

Conclusions:

  • Integrating artifact-aware preprocessing with spatial-contextual feature learning significantly improves forgery detection reliability.
  • This work establishes a foundation for a new generation of trustworthy digital media verification systems.
  • Challenges remain in detecting very small manipulations and heavily compressed images.