MesoNet: a Compact Facial Video Forgery Detection Network
Darius Afchar, Vincent Nozick, Junichi Yamagishi, Isao Echizen
Code Available — Be the first to reproduce this paper.
ReproduceCode
- github.com/DariusAf/MesoNetOfficialIn papertf★ 0
- github.com/Recognito-Vision/Linux-FaceRecognition-FaceLivenessDetectionnone★ 363
- github.com/FaceOnLive/DeepFake-Detection-SDK-Linuxnone★ 59
- github.com/tamlhp/dfd_benchmarkpytorch★ 23
- github.com/HongguLiu/MesoNet.Pytorchpytorch★ 0
- github.com/MalayAgarwal-Lee/MesoNet-DeepFakeDetectiontf★ 0
- github.com/maragori/DeepfakeForensics-v1pytorch★ 0
- github.com/Raj-08/Deepfake-Detection-Mesonettf★ 0
Abstract
This paper presents a method to automatically and efficiently detect face tampering in videos, and particularly focuses on two recent techniques used to generate hyper-realistic forged videos: Deepfake and Face2Face. Traditional image forensics techniques are usually not well suited to videos due to the compression that strongly degrades the data. Thus, this paper follows a deep learning approach and presents two networks, both with a low number of layers to focus on the mesoscopic properties of images. We evaluate those fast networks on both an existing dataset and a dataset we have constituted from online videos. The tests demonstrate a very successful detection rate with more than 98% for Deepfake and 95% for Face2Face.