Cavernous Hoax Scanner is a web-based platform that leverages JPEN stack technology to implement deep learning algorithms, specifically CNN, GANs, Relational regression.

Code, model and data for our paper: K. Tsigos, E. Apostolidis, S. Baxevanakis, S. Papadopoulos, V. Mezaris, "Towards Quantitative Evaluation of Explainable AI Methods for Deepfake Detection", Proc. ACM Int. Workshop on Multimedia AI against Disinformation (MAD’24) at the ACM Int. Conf. on Multimedia Retrieval (ICMR’24), Thailand, June 2024.

[NeurIPS 2024] This is the official repo of the paper "Lips Are Lying: Spotting the Temporal Inconsistency between Audio and Visual in Lip-syncing DeepFakes".

A Survey on Deepfake Generation and Detection

Paid Project

Real-Time Deepfake Detection in the Real-World

**VeriVideo** is a deepfake face detection system designed to ensure digital media integrity. Using ResNeXt-50, LSTM, and fully connected layers, it achieves 95.22% accuracy. Integrated into a scalable web app, VeriVideo empowers users to detect manipulated videos, fostering trust in digital content.

Face Liveness Detection SDK - iBeta level 2 compliant passive face liveness detection SDK which can detect printed photos, video replay, 3D masks, and deepfake threats

Liveness detection SDK Linux - iBeta level 2 compliant 3D passive liveness detection engine which can detect printed photos, video replay, 3D masks, and deepfake threats

Liveness detection SDK iOS- iBeta level 2 certified 3D passive liveness detection engine which can detect printed photos, video replay, 3D masks, and deepfake threats

Liveness detection SDK Android - iBeta level 2 compliant 3D passive liveness detection engine which can detect printed photos, video replay, 3D masks, and deepfake threats

Final year project on Deep-Fake Detection

Code and pre-trained models for our paper "CLIPping the Deception: Adapting Vision-Language Models for Universal Deepfake Detection".

Classic BWS: Liveness Detection, Deepfake Detection and PhotoVerify Face Match for KYC

3D Passive Face Liveness Detection (Anti-Spoofing) & Deepfake detection. A single image is needed to compute liveness score. 99,67% accuracy on our dataset and perfect scores on multiple public datasets (NUAA, CASIA FASD, MSU...).

A list of tools, papers and code related to Deepfake Detection.

On-Premise DeepFake Detection SDK for Linux

[ICASSP 2024] The official repo for Harnessing the Power of Large Vision Language Models for Synthetic Image Detection

本科毕业设计：针对Deepfake假脸视频面部细节特征的提取算法

Countermeasures for Misinformation & Plug-In for Detection by Team AceDefenders.