Synthetic Aperture Radar (SAR) images are a valuable asset for a wide variety of tasks. In the last few years, many websites have been offering them for free in the form of easy to manage products, favoring their widespread diffusion and research work in the SAR field. The drawback of these opportunities is that such images might be exposed to forgeries and manipulations by malicious users, raising new concerns about their integrity and trustworthiness. Up to now, the multimedia forensics literature has proposed various techniques to localize manipulations in natural photographs, but the same problem has never been investigated on SAR images. Forensics methods developed for natural photographs are not guaranteed to succeed on SAR images, as their generation pipeline is completely different from that of digital pictures. In this paper, we investigate the problem of localizing splicing attacks in amplitude SAR imagery. Our goal is to identify the pixels of an amplitude SAR image that have been copied and pasted from another image for malicious purposes, considering also that the attacker might have applied some editing to conceal this manipulation. To do so, we leverage a Convolutional Neural Network (CNN) to extract a fingerprint highlighting inconsistencies in the processing traces of the analyzed input. Then, we examine this fingerprint to produce a binary tampering mask indicating the pixel region under splicing attack. Results show that our proposed method, tailored to the nature of SAR signals, provides better performances than state-of-the-art forensic tools developed for natural images.

Amplitude SAR Imagery Splicing Localization

Cannas, Edoardo Daniele;Bonettini, Nicolo;Mandelli, Sara;Bestagini, Paolo;Tubaro, Stefano
2022

Abstract

Synthetic Aperture Radar (SAR) images are a valuable asset for a wide variety of tasks. In the last few years, many websites have been offering them for free in the form of easy to manage products, favoring their widespread diffusion and research work in the SAR field. The drawback of these opportunities is that such images might be exposed to forgeries and manipulations by malicious users, raising new concerns about their integrity and trustworthiness. Up to now, the multimedia forensics literature has proposed various techniques to localize manipulations in natural photographs, but the same problem has never been investigated on SAR images. Forensics methods developed for natural photographs are not guaranteed to succeed on SAR images, as their generation pipeline is completely different from that of digital pictures. In this paper, we investigate the problem of localizing splicing attacks in amplitude SAR imagery. Our goal is to identify the pixels of an amplitude SAR image that have been copied and pasted from another image for malicious purposes, considering also that the attacker might have applied some editing to conceal this manipulation. To do so, we leverage a Convolutional Neural Network (CNN) to extract a fingerprint highlighting inconsistencies in the processing traces of the analyzed input. Then, we examine this fingerprint to produce a binary tampering mask indicating the pixel region under splicing attack. Results show that our proposed method, tailored to the nature of SAR signals, provides better performances than state-of-the-art forensic tools developed for natural images.
SAR , GRD , image splicing localization , deep learning , multimedia forensics , satellite imagery
File in questo prodotto:
File Dimensione Formato  
ACCESS3161836.pdf

accesso aperto

Descrizione: Author Accepted Manuscript
: Publisher’s version
Dimensione 3.84 MB
Formato Adobe PDF
3.84 MB Adobe PDF Visualizza/Apri

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: http://hdl.handle.net/11311/1208027
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 0
  • ???jsp.display-item.citation.isi??? ND
social impact