A survey of machine learning techniques in adversarial image forensics

Nowroozi, Ehsan; Dehghantanha, Ali; Parizi, Reza M.; Choo, Kim‐Kwang Raymond

doi:10.1016/j.cose.2020.102092

Cited by 58 publications

(27 citation statements)

References 51 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This is a very challenging problem, since several artifacts, such as the correlation between RGB channels, discrete cosine transform irregularities, and even illumination inconsistencies in the digital image versions are usually removed by the print and scan process. In this way, although several adversarial attacks have been discussed in the literature [ 55 ], the print and scan procedure is the easiest yet most powerful attack an adversary could perform against deepfake digital image detectors.…”

Section: Related Workmentioning

confidence: 99%

VIPPrint: Validating Synthetic Image Detection and Source Linking Methods on a Large Scale Dataset of Printed Documents

2021

Self Cite

View full text Add to dashboard Cite

The possibility of carrying out a meaningful forensic analysis on printed and scanned images plays a major role in many applications. First of all, printed documents are often associated with criminal activities, such as terrorist plans, child pornography, and even fake packages. Additionally, printing and scanning can be used to hide the traces of image manipulation or the synthetic nature of images, since the artifacts commonly found in manipulated and synthetic images are gone after the images are printed and scanned. A problem hindering research in this area is the lack of large scale reference datasets to be used for algorithm development and benchmarking. Motivated by this issue, we present a new dataset composed of a large number of synthetic and natural printed face images. To highlight the difficulties associated with the analysis of the images of the dataset, we carried out an extensive set of experiments comparing several printer attribution methods. We also verified that state-of-the-art methods to distinguish natural and synthetic face images fail when applied to print and scanned images. We envision that the availability of the new dataset and the preliminary experiments we carried out will motivate and facilitate further research in this area.

show abstract

Section: Related Workmentioning

confidence: 99%

VIPPrint: Validating Synthetic Image Detection and Source Linking Methods on a Large Scale Dataset of Printed Documents

2021

Self Cite

View full text Add to dashboard Cite

show abstract

“…Another set of surveys focus on the specific context of child abuse material and its detection through image and video analysis [111]- [114]. More recently, the advent of deep learning techniques has enhanced the capabilities of image integrity detection and verification, outperforming traditional methods in several image-related tasks, especially in these where anti-forensic tools were used [109], [110], [115]. In the context of video files, we can find surveys on video steganalysis [109], [110], [116], video forgery detection [91], [92], [94], [110], [117], [118], video forensic tools [91], [109], [119], [120], video surveillance analysis [121], [122], and video content authentication [123].…”

Section: Challenge/limitation Referencesmentioning

confidence: 99%

Research Trends, Challenges, and Emerging Topics in Digital Forensics: A Review of Reviews

et al. 2022

View full text Add to dashboard Cite

This work was supported by the European Commission under the Horizon 2020 Programme (H2020), as part of the projects LOCARD (https://locard.eu) (Grant Agreement no. 832735) and CyberSec4Europe (https://www.cybersec4europe.eu) (Grant Agreement no. 830929). F. Casino was supported by the Beatriu de Pinós programme of the Government of Catalonia (Grant No. 2020 BP 00035). The content of this article does not reflect the official opinion of the European Union. Responsibility for the information and views expressed therein lies entirely with the authors.

show abstract

“…The problem of detecting privacy enhancement, on the other hand, has not yet been explored in the literature. Because some privacy models are based on adversarial perturbations, this problem is also partially related to adversarial-attack detection techniques [36][37][38][39][40]. However, since soft-biometric privacy enhancement also includes synthesis-based methods, data hidding solutions and a wide variety of other approaches [10], the problem of detecting such image modifications is considerably broader.…”

Section: Detecting Privacy Enhancementmentioning

confidence: 99%

Detecting Soft-Biometric Privacy Enhancement

Rot

Peer

2022

Handbook of Digital Face Manipulation and Detection

View full text Add to dashboard Cite

With the proliferation of facial analytics and automatic recognition technology that can automatically extract a broad range of attributes from facial images, so-called soft-biometric privacy-enhancing techniques have seen increased interest from the computer vision community recently. Such techniques aim to suppress information on certain soft-biometric attributes (e.g., age, gender, ethnicity) in facial images and make unsolicited processing of the facial data infeasible. However, because the level of privacy protection ensured by these methods depends to a significant extent on the fact that privacy-enhanced images are processed in the same way as non-tampered images (and not treated differently), it is critical to understand whether privacy-enhancing manipulations can be detected automatically. To explore this issue, we design a novel approach for the detection of privacy-enhanced images in this chapter and study its performance with facial images processed by three recent privacy models. The proposed detection approach is based on a dedicated attribute recovery procedure that first tries to restore suppressed soft-biometric information and based on the result of the restoration procedure then infers whether a given probe image is privacy enhanced or not. It exploits the fact that a selected attribute classifier generates different attribute predictions when applied to the privacy-enhanced and attribute-recovered facial images. This prediction mismatch (PREM) is, therefore, used as a measure of privacy enhancement. In extensive experiments with three popular face datasets we show that the proposed PREM model is able to accurately detect privacy enhancement in facial images despite the fact that the technique requires no supervision, i.e., no examples of privacy-enhanced images are needed for training.

show abstract

A survey of machine learning techniques in adversarial image forensics

Cited by 58 publications

References 51 publications

VIPPrint: Validating Synthetic Image Detection and Source Linking Methods on a Large Scale Dataset of Printed Documents

VIPPrint: Validating Synthetic Image Detection and Source Linking Methods on a Large Scale Dataset of Printed Documents

Research Trends, Challenges, and Emerging Topics in Digital Forensics: A Review of Reviews

Detecting Soft-Biometric Privacy Enhancement

Contact Info

Product

Resources

About