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.

https://hal-upec-upem.archives-ouvertes.fr/hal-01867298/document

MesoNet: a Compact Facial Video Forgery Detection Network

MesoNet: a Compact Facial Video Forgery Detection Network.

Validating a given multimedia content is nowadays quite a hard task because of the huge amount of possible alterations that could have been operated on it. In order to face this problem, image and video experts have proposed a wide set of solutions to reconstruct the processing history of a given multimedia signal. These strategies rely on the fact that non-reversible operations applied to a signal leave some traces ("footprints") that can be identified and classified in order to reconstruct the possible alterations that have been operated on the original source. These solutions permit also to identify which source generated a specific image or video content given some device-related peculiarities. The paper aims at providing an overview of the existing video processing techniques, considering all the possible alterations that can be operated on a single signal and also the possibility of identifying the traces that could reveal important information about its origin and use.

/pdf/an-overview-on-video-forensics-20693kkcsx.pdf

An overview on video forensics

Ternary radial harmonic Fourier moments based robust stereo image zero-watermarking algorithm

In this paper we will present the concept of a system that allows for an evolutionary introduction of depth perception into the existing 2D digital TV framework. The work is part of the European Information Society Technologies (IST) project “Advanced Three-Dimensional Television System Technologies” (ATTEST), an activity where industries, research centers and universities have joined forces to design a backwardscompatible, flexible and modular broadcast 3D-TV system, where all parts of the 3D processing chain are optimised to one another. This includes content creation, coding and transmission, display and research in human 3D perception, which be will used to guide the development process. The goals of the project comprise the development of a novel broadcast 3D camera, algorithms to convert existing 2D-video material into 3D, a 2Dcompatible coding and transmission scheme for 3D-video using MPEG2/4/7 technologies and the design of two new autostereoscopic displays.

IST project ATTEST - an evolutionary and optimised approach on 3D-TV

The popularity of 3D content is on the rise since it provides an immersive experience to viewers. In this situation, depth-image-based rendering (DIBR) has taken on an important role in 3D technology due to its low bandwidth cost and ease of depth configuration. Noting that the viewer could record provided center view or synthesized views for illegal distribution, it is clear that copyright protection must be taken into account for the DIBR 3D content, including the possibility that one single view could be illegally distributed as 2D content. In this paper, we propose a robust watermarking scheme for DIBR 3D images by quantization on dual-tree complex wavelet transform (DT-CWT) coefficients with consideration of imperceptibility. To make the proposed scheme robust to DIBR process, two characteristics of DT-CWT are employed: approximate shift invariance and directional selectivity. We select certain coefficient sub-blocks and group the coefficient rows based on the properties of DIBR. On the extraction side, the threshold is carefully chosen with a low false positive rate. The simulation results show that the embedded watermark is stably extracted from the center view and the synthesized left and right views. In addition, even if the synthesized left and right views are distorted by general attacks, the watermark is successfully extracted. Furthermore, the proposed scheme is robust to pre-processing of the depth image and baseline adjusting, which are common processing on the DIBR system for better quality of 3D views.

Robust DT-CWT Watermarking for DIBR 3D Images

Region-based tampering detection and recovery using homogeneity analysis in quality-sensitive imaging

In this paper, we propose a blind watermark detection scheme for digital cinema which is robust against D-A/A-D conversion attack. D-A/A-D conversion attack commonly happens during capturing cinema with a camera and is composed of various geometrical distortions as well as signal processing distortions. In many video applications, geometric distortions have been compensated by utilizing auto-correlation function (ACF). Since the ACF cannot resist projective transform but affine transform, we present a robust watermark detection scheme utilizing local auto-correlation function (LACF) with a mathematical model. Also, we suggest a way to design watermark pattern to improve performance of the LACF. We demonstrate robustness of the presented scheme against D-A/A- D conversion attacks for digital cinema as well as common video processing attacks.

https://hklee.kaist.ac.kr/publications/2008%20ICIP%20(with%20Lee,%20Min-Jeong).pdf

Robust watermark detection against D-A/A-D conversion for digital cinema using local auto-correlation function

As screenshots of copyrighted video contents are spreading through the Internet without any regulation, the copyright infringement arises gradually. However, there are still no techniques that prevent illegal screenshots from spreading out. Since most television systems use interlaced scanning mode and camcorders offer interlaced recording, many screenshots are interlaced. In this paper, we propose a screenshot identification scheme using unique characteristic of interlaced video, combing artifact. To do this, several blocks of the input image are selected for the significant combing artifact they have. In each block, eight features that represent the artifact are extracted. These extracted features are applied to train and test support vector machine for identifying whether the input image is a screenshot or not. In the experiments, the identification accuracy is measured when the input image is converted from and to common multimedia formats. The results prove that the proposed identification scheme performs well under the widely used image formats as JPEG, BMP, TIFF and video formats as MPEG-2, MPEG-4, H.264.

http://hklee.kaist.ac.kr/publications/2010%20ACM%20MMSec(with%20Lee,%20Ji-Won).pdf

Screenshot identification using combing artifact from interlaced video

There exist various video watermarking schemes to protect copyrighted video content. Since most of these schemes exploit the perceptual properties of an original video content, they ineludibly had high computing complexity to embed watermark. However, some video applications such as settop box (STB) don't have enough computing power to insert the watermark into the high definition (HD) content using existing schemes. In this study, we propose an overlay video watermarking scheme that embeds watermark into the overlay layer of the multimedia device. Since the proposed scheme exploits only the traits of human visual system (HVS) without any information about original video data, it achieves both low computational complexity and good imperceptibility. Further, the experimental results prove that the proposed scheme has good fidelity and enough robustness.

http://hklee.kaist.ac.kr/publications/2011%20IIH-MSP%28with%20Ji-Won%20Lee%29.pdf?origin=publication_detail

Tae-Woo Oh

Papers

Robust DT-CWT Watermarking for DIBR 3D Images

Region-based tampering detection and recovery using homogeneity analysis in quality-sensitive imaging

Robust watermark detection against D-A/A-D conversion for digital cinema using local auto-correlation function

Screenshot identification using combing artifact from interlaced video

Video Watermarking on Overlay Layer