Forgery Detection by Internal Positional Learning of Demosaicing Traces

Quentin Bammey, Rafael Grompone von Gioi, Jean-Michel Morel

Research output: Chapters, Conference Papers, Creative and Literary WorksRGC 32 - Refereed conference paper (with host publication)peer-review

9 Citations (Scopus)

Abstract

We propose 4Point (Forensics with Positional Internal Training), an unsupervised neural network trained to assess the consistency of the image colour mosaic to find forgeries. Positional learning trains the model to learn the modulo-2 position of pixels, leveraging the translation-invariance of CNN to replicate the underlying mosaic and its potential inconsistencies. Internal learning on a single potentially forged image improves adaption and robustness to varied post-processing and counter-forensics measures. This solution beats existing mosaic detection methods, is more robust to various post-processing and counter-forensic artefacts such as JPEG compression, and can exploit traces to which state-of-the-art generic neural networks are blind. Check qbammey.github.io/4point for the code. © 2022 IEEE.
Original languageEnglish
Title of host publicationProceedings - 2022 IEEE/CVF Winter Conference on Applications of Computer Vision, WACV 2022
PublisherIEEE
Pages1019-1029
ISBN (Electronic)9781665409155
ISBN (Print)9781665409162
DOIs
Publication statusPublished - 2022
Externally publishedYes
Event22nd IEEE/CVF Winter Conference on Applications of Computer Vision (WACV 2022) - Waikoloa, United States
Duration: 4 Jan 20228 Jan 2022

Publication series

NameProceedings - IEEE/CVF Winter Conference on Applications of Computer Vision, WACV
ISSN (Print)2472-6737
ISSN (Electronic)2642-9381

Conference

Conference22nd IEEE/CVF Winter Conference on Applications of Computer Vision (WACV 2022)
PlaceUnited States
CityWaikoloa
Period4/01/228/01/22

Research Keywords

  • Few-shot
  • Semi- and Un- supervised Learning Image forensics
  • Transfer

Fingerprint

Dive into the research topics of 'Forgery Detection by Internal Positional Learning of Demosaicing Traces'. Together they form a unique fingerprint.

Cite this