synthetic-media 开源项目推荐（50 个）

aaronchong888 /DeepFake-Detect

Open-source deepfake detection: train your own model with TensorFlow, Keras & EfficientNet

resemble-ai /detect-skill

Agent skill for deepfake detection & media safety — detect AI-generated audio, images, and video with Resemble AI

bes-dev /pytorch_clip_guided_loss

A simple library that implements CLIP guided loss in PyTorch.

Awaisali36 /ai-avatar-video-generation-system

🎬 Automated AI avatar news video generator built with n8n. Fetches RSS headlines, generates scripts with Google Gemini, and creates professional videos with HeyGen's AI avatars. Fully automated from feed to final video.

moritztng /cyberstudio

Face Animation from Text 🧙‍♂️

OmerKurtulus /DeepReality

PIN Architecture: a six-layer forensic system for detecting AI-generated and manipulated imagery. Fifteen independent analysis pins run concurrently: C2PA provenance, compression forensics, CLIP/SigLIP2/frequency detectors, Grad-CAM explainability, LLM adjudication and a calibrated XGBoost ensemble. Signed provenance outranks statistical inference.

thecosmicskye /open-persona-license

Open Persona License — A public license for persona, identity, and synthetic-media reuse.

iblai /video

AI video studio on the ibl.ai platform for generating avatar videos, cloning voices, and authoring scripts from a single workspace. Built with Next.js, TypeScript, and Tailwind CSS on the @iblai/iblai-js SDK, with a hosted backend at iblai.app. Runs on desktop and mobile.

tattle-made /deepfake-marker

Open Source Deepfake Detection Models in One Place

devghori1264 /Aletheia

Aletheia is a multimodal ML architecture for forensic detection of synthetic video. Integrating deep learning and NLP, it analyzes spatiotemporal anomalies, biometric artifacts, and audio-linguistic misalignments. The framework performs cross-modal feature extraction to provide deterministic probability scoring for media authentication.

mfs-iplab /aadd-2025

Official repository for AADD-2025: 1st Adversarial Attacks on Deepfake Detectors Challenge - ACM Multimedia 2025 Grand Challenge. Investigating adversarial vulnerabilities in deepfake detection models through adversarial perturbations that evade classifiers while preserving visual similarity.

thecosmicskye /skye-open-persona

Skye - Open Persona — Source media and persona reference material for Skye under the Open Persona License.

ishandutta2007 /Awesome-Deepfake-Detection

该项目在 GitHub 上没有填写描述

aloth /verification-crisis

Expert survey on GenAI disinformation threats & countermeasures - instrument, data, and publications. Published in HKS Misinformation Review (doi:10.37016/mr-2020-205) and at R2CASS @ WWW 2026. Survey still open for expert participants.

xayakster /Deeptrace-AI

A tri-branch digital media forensics workstation for detecting AI-generated synthetic images (Diffusion/GANs) and Photoshop tampering with calibrated Conformal Prediction guarantees.

umitkacar /synthetic-media-challenge-solutions

DeepFake detection with Deep Tree Network (DTN) architecture - 94.5% accuracy, 45 FPS real-time processing with zero-shot learning

Omega-Point-Solutions /omega-lens-public

Omega Lens: sovereign deepfake and synthetic-media recognition + context restoration. By Omega Point Solutions LLC (Veteran-Owned Small Business). Public showcase, determination charter, and REST SDK clients.

sreenathyadavk /HiddenLayer

Real-time deepfake detection Android app using multi-modal AI fusion. Features CNN inference, artifact analysis & on-device ML. Built with Kotlin, Jetpack Compose & TensorFlow Lite.

isfakeai /bbc-ai-content-analysis

Analysis of 13 years of BBC articles (2012–2025) using the isFake.ai AI text detector.

diegormirhan /prooflens

Privacy-first image authenticity workbench combining C2PA provenance, digital forensics, metadata analysis, and local AI-generated image screening with React, FastAPI, and ONNX.