Exposing digital forgeries by detecting traces of resampling

By Alin C. Popescu, Hany Farid. 2005

1,029 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Safety, Robustness & Interpretability

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Safety, Robustness & Interpretability

Citations in the corpus

CNN-Generated Images Are Surprisingly Easy to Spot… for Now (2019, background): “Early work focused on hand-crafted cues [14] such as compression artifacts [3], resampling [31], or physical scene constraints [27].”
Detecting digital image forgeries using sensor pattern noise (2006, background): “Because these low frequency components are not a characteristic of the sensor, we do not use them for identification of forgeries.”
Digital Image Authentication From JPEG Headers (2011, background): “tampering: region duplication [1], [2], [3], [4]; resampling [5], [6]; color filter array artifacts [7], [8]; inconsistencies in camera response function [9]; inconsistencies in lighting and shadows [10], [11], [12], [13]; inconsistencies in chromatic aberrations [14], [15]; inconsistencies in senso”
Using noise inconsistencies for blind image forensics (2009, background): “An ROI can be identified by one of the forgery detection techniques capable of localizing the tampering as well (for example [4,10, 6 ]).”
Digital image forensics via intrinsic fingerprints (2008, background): “Thus, post-processing oper ations such as resampling can be identified by studying the induced correlations [ 15 ].”
Median Filtering Forensics Based on Convolutional Neural Networks (2015, background): “Recently more efforts have been made to expose the processing history of digital images, such as filtering [1]–[5], re-sampling [6], compression [7], and contrast enhancement [8], since their blind detections are forensically important [1]–[5].”
Constrained Convolutional Neural Networks: A New Approach Towards General Purpose Image Manipulation Detection (2018, methodology): “Many forensic algorithms, such as resampling detectors [2], [3], median filtering detectors [7] and other forensic detectors based on steganalytic features like the SRM [20], operate by first extracting prediction residual features, then by forming higher-level features from these residuals.”
A Deep Learning Approach to Universal Image Manipulation Detection Using a New Convolutional Layer (2016, methodology): “This approach has been used to develop algorithms targeted at detecting image manipulations such as resizing and resampling [17, 8], median ﬁltering [9, 7], contrast enhancement [19], etc.”
Source Camera Identification Using Enhanced Sensor Pattern Noise (2009, background): “…source devices, classifying images, and verifying the integrity of images based on the detection of existence or local inconsistencies of device attributes or data processing related characteristics, such as sensor pattern noise (SPN) [1]–[8], camera response function [9], resampling artefacts…”
Information Forensics: An Overview of the First Decade (2013, background): “Contrast enhancement can be detected by measuring the strength of these fingerprints in an image’s pixel value histogram [22].”
Exposing Digital Forgeries From JPEG Ghosts (2009, methodology): “These include techniques for detecting cloning [1], [2]; splicing [3]; re-sampling artifacts [4], [ 5 ] ; color filter array aberrations [6]; disturbances of a camera’s sensor noise pattern [7]; chromatic aberrations [8]; and lighting inconsistencies [9], [10], [11].”
What makes fake images detectable? Understanding properties that generalize (2020, background): “A possible solution involves checking for consistency throughout the image – examples include predicting metadata [14] or other low-level artifacts [27,28,29], learning similar embeddings for nearby patches [38], or learning similarity graphs from image patches [25].”
Exposing DeepFake Videos By Detecting Face Warping Artifacts (2018, methodology): “n essence due to afﬁne transforms to the synthesized face. In the literature of digital media forensics, detecting transforms or the underlying resampling algorithm has been extensively studied, e.g, [15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25]. However, the performance of these methods are affected”
Exposing Digital Forgeries in Complex Lighting Environments (2007, methodology): “Statistical techniques have been developed for detecting cloning [1], [2]; splicing [3]; resampling artifacts [4], [5]; color filter array aberrations [4]; and disturbances of a camera’s sensor noise pattern [6].”
Detecting Digital Image Forgeries by Measuring Inconsistencies of Blocking Artifact (2007, background): “The first attempt is to detect the traces of specific manipulation during forgery creation, such as resampling [1], copypaste [2, 3], and color filter array interpolation [4].”
Image forgery detection (2009, background): “Because these correlations are unlikely to occur naturally, their presence can be used to detect this specific manipulation ([36] ; related approaches are described in [43],[22], [31], and [38] ).”
Fighting Fake News: Image Splice Detection via Learned Self-Consistency (2018, methodology): “Drawing upon techniques from signal processing, previous methods focused on cues such as misaligned JPEG blocks [10], compression quantization artifacts [11], resampling artifacts [12], color filtering array discrepancies [13], and camera-hardware “fingerprints” [14].”
Recasting Residual-based Local Descriptors as Convolutional Neural Networks: an Application to Image Forgery Detection (2017, background): “A number of papers have been proposed to detect one or the other of such elaborations [13, 15, 16, 22, 25].”
Exposing digital forgeries in color filter array interpolated images (2005, methodology): “In contrast to these approaches, we and others have proposed techniques for detecting tampering in digital images that work in the absence of any digital watermark or signature [4], [5], [6], [7], [ 8 ], [9], [10].”