NoiseScope: Detecting Deepfake Images in a Blind Setting

Pu, Jiameng; Mangaokar, Neal; Wang, Bolun; Reddy, Chandan K.; Viswanath, Bimal

doi:10.1145/3427228.3427285

Cited by 21 publications

(8 citation statements)

References 49 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Rodriguez et al [57] used this noise pattern in detecting the deepfake. Pu et al [58] proposed a method for deepfake detection called it noise scope which consists of four main parts: noise extractor, fingerprint extractor, fingerprint classifier, and finally fake image detector. This method detects deepfake in a blind way i.e.…”

Section: Multimedia Forensics Based Methodsmentioning

confidence: 99%

A survey of deepfakes in terms of deep learning and multimedia forensics

Hadi

Kadhem²,

Abbas³

2022

IJECE

View full text Add to dashboard Cite

Artificial intelligence techniques are reaching us in several forms, some of which are useful but can be exploited in a way that harms us. One of these forms is called deepfakes. Deepfakes is used to completely modify video (or image) content to display something that was not in it originally. The danger of deepfake technology impact on society through the loss of confidence in everything is published. Therefore, in this paper, we focus on deepfakedetection technology from the view of two concepts which are deep learning and forensic tools. The purpose of this survey is to give the reader a deeper overview of i) the environment of deepfake creation and detection, ii) how deep learning and forensic tools contributed to the detection of deepfakes, and iii) finally how in the future incorporating both deep learning technology and tools for forensics can increase the efficiency of deepfakes detection.

show abstract

Section: Multimedia Forensics Based Methodsmentioning

confidence: 99%

A survey of deepfakes in terms of deep learning and multimedia forensics

Hadi

Kadhem²,

Abbas³

2022

IJECE

View full text Add to dashboard Cite

show abstract

“…[ 39 ] proposes an unsupervised approach to detecting GAN(deepfake) images. This approach works even without access to deepfake images during training, which makes it highly generalizable.…”

Section: Deepfakesmentioning

confidence: 99%

“…One of the things mentioned is that similarly to camera fingerprinting, GAN model image generation leaves unique noise patterns related to the model that gener-ated the image. The model extracts this fingerprint to detect GAN images and is agnostic to the type of GAN used to generate the image [ 39 ]. An example of different types of fingerprints can be seen in Fig.…”

Section: Deepfakesmentioning

confidence: 99%

“…These will return in any picture created with that camera or generated with that GAN. The proposed method should me more useful than supervised training methods as new GAN models are proposed very often, leav-ing supervised methods behind in detecting the new types of GAN images [ 39 ]. shows the difference between camera fingerprints, also known as photo-response non-uniformity (PRNU) patterns and images generated by GAN models.…”

Section: Deepfakesmentioning

confidence: 99%

See 1 more Smart Citation

Interpol review of forensic video analysis, 2019–2022

Geradts

Riphagen

2023

Forensic Science International: Synergy

View full text Add to dashboard Cite

“…Deepfakes predominantly represent research by academics and industries in the specific fields of computer vision and ML/AI in computer science. They mostly focus on projects for improving deepfake techniques [9,13,55,57] and methods for detecting deepfakes [1,33,34,46,67]. A recent survey on deepfake creation and detection techniques predicts that deepfakes can be weaponized for the monetization of its products and services [41].…”

Section: Deepfake Technologymentioning

confidence: 99%

Are Deepfakes Concerning? Analyzing Conversations of Deepfakes on Reddit and Exploring Societal Implications

Gamage,

Ghasiya,

Bonagiri

et al. 2022

Preprint

View full text Add to dashboard Cite

Deepfakes are synthetic content generated using advanced deep learning and AI technologies. The advancement of technology has created opportunities for anyone to create and share deepfakes Permission to make digital or hard copies of part or all of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. Copyrights for third-party components of this work must be honored. For all other uses, contact the owner/author(s).

show abstract

NoiseScope: Detecting Deepfake Images in a Blind Setting

Cited by 21 publications

References 49 publications

A survey of deepfakes in terms of deep learning and multimedia forensics

A survey of deepfakes in terms of deep learning and multimedia forensics

Interpol review of forensic video analysis, 2019–2022

Are Deepfakes Concerning? Analyzing Conversations of Deepfakes on Reddit and Exploring Societal Implications

Contact Info

Product

Resources

About