video-forensics

EfficientNet+MTCNN deepfake detection achieving 87% accuracy with 12.96% EER - Multi-modal video forensics for misinformation prevention

Next-gen video platform.

HLS Stream Analyzer – A passive, security-focused tool to analyze and reconstruct HLS media streams, helping researchers identify exposed segments, misconfigurations, and potential risks in streaming infrastructure.

Real-event-anchored benchmark for detecting AI-generated videos in real-world crisis settings.

A deepfake detection platform using fine-tuned EfficientNet-B4. Features GradCAM explainability to show exactly where a video was manipulated, alongside an InsightFace pipeline.

fractalVideoGuard

aI-powered multimedia forensic analysis system for detecting CCTV tampering, deepfakes, and metadata manipulation.

Frequency-aware deepfake detection under lossy H.264 compression using XceptionNet, DCT features, learnable frequency masking, and cross-attention fusion.

A tool that turns live cameras into a real-time threat-detection and search command center

Training-free synthetic-video detection via second-order motion. Found the published 100%/54pp is close to best-case for the injected artifact shape -- a less tailored, equally plausible artifact drops it to ~73%/23pp, still a real but more modest effect. Verified on tuning+holdout.

Cryptographic proof of video authenticity. Verify video origin and integrity using perceptual hashing and digital signatures to combat manipulated media.

SaksiAI is an enterprise-grade Evidence Intelligence Platform (EIP) specifically designed for the Indonesian market. It addresses the Hidden Data Problem where critical organizational knowledge is trapped in unusable, fragmented formats such as scanned PDFs, voice recordings, and CCTV feeds.

Public-safe research snapshot for deepfake detection, temporal localization, and ongoing audio-visual investigation.

Evidence-first diagnostics for video artifacts, file lineage, encoding, delivery, and enhancement

Full-stack deepfake video detection system — EfficientNetV2-B3 + Temporal Attention BiLSTM | 99.54% AUC on Celeb-DF | React frontend + FastAPI backend

Production-grade deepfake detection — 3-model ensemble (Xception + EfficientNet-B4 + ViT), calibrated confidence, forensics suite, REST API, and video AV-sync/identity tracking.

A low-latency deepfake video detection pipeline utilizing frequency-domain analysis (FFT) and a fine-tuned ResNet50 for rapid forensic triage.

Deepfake video detection using MobileNet and MesoNet CNNs with LSTM for temporal feature analysis across FaceForensics++, Celeb-DF, and DFDC datasets.

Video forensics for deepfakes, AI-generated content, and miscontextualized footage

Official documentation and technical guides for VideoJIN Deepfake Detection API & Media Forensics.