Skip to Main content Skip to Navigation

Développement d'une technique robuste du tatouage des images aux attaques print-cam

Abstract : Digital image watermarking consists in embedding information within the image which cannot be seen by the human visual system, but recovered with a software. The aim of this thesis is to propose a watermarking method when the watermarked image is printed on a physical support and then read freehandedly with a smartphone camera. In order to survive to the print-cam process, the watermark must resist to multiple attacks. Those attacks occur during printing and capturing the image with a camera: in that case the captured image might be rotated around the optical axis of the camera producing 3D geometric deformation and translations and scaling may also occur. Pixel value distortions are also present. These attacks may cause the loss of synchronization of the watermark and make the detection impossible. Hence, the main objective of the thesis is to develop a watermarking method that is robust to the print-cam attack in the context of an industrial security application for ID images. Fourier transform is used as this watermarking domain has invariance properties against some geometric distortions. Three main correction methods were integrated to deal with print-cam attacks: frame-based perspective rectification of the freehandedly captured images using detection of Hough lines, a Wiener filter to decrease image blurring and reduce noise, and adjustments to reduce color degradations. Results show that the method is highly robust with a total error rate of 1%, compared to a least 25% of errors for other methods. the obtained error rate (1%) is being compatible with the targeted industrial application.
Document type :
Complete list of metadata
Contributor : Abes Star :  Contact Connect in order to contact the contributor
Submitted on : Tuesday, February 16, 2021 - 1:58:08 PM
Last modification on : Tuesday, October 12, 2021 - 5:20:34 PM
Long-term archiving on: : Monday, May 17, 2021 - 7:39:44 PM


Version validated by the jury (STAR)


  • HAL Id : tel-03142938, version 1


Khadija Gourrame. Développement d'une technique robuste du tatouage des images aux attaques print-cam. Autre [cs.OH]. Université d'Orléans; Université Ibn Zohr (Agadir), 2019. Français. ⟨NNT : 2019ORLE3047⟩. ⟨tel-03142938⟩



Record views


Files downloads