DFT-based watermarking method for medical images

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

6 Scopus citations

Abstract

We present a robust DFT-based watermarking method applied to medical imaging stored in DICOM standard format. The watermark is a 1-D binary pseudo-random pattern generated by a secret key which is embedded into the magnitude of the DFT middle frequencies of the original image in order to preserve a high quality perception. To increase security of the watermark, data bits are scrambled using a secret key. During the extraction stage, watermark data bits are extracted, re-arranged with the same secret key used in the embedding process and detected using the bit error rate (BER) criteria. The watermarked image quality is measured using the Peak Signal to Noise Ratio (PSNR) and Structural Similarity Index (SSIM). The experimental results show the watermark robustness against Rotation, Scaling and Translation (RST) geometric distortions and some common signal processing operations, including DICOM compression schemes, additive Gaussian and impulsive noise, Gaussian and median filtering, sharpness, contrast and brightness changes.

Original languageEnglish
Title of host publicationProceedings - 2012 9th Electronics, Robotics and Automotive Mechanics Conference, CERMA 2012
Pages21-24
Number of pages4
DOIs
StatePublished - 2012
Event2012 9th Electronics, Robotics and Automotive Mechanics Conference, CERMA 2012 - Cuernavaca, Morelos, Mexico
Duration: 19 Nov 201223 Nov 2012

Publication series

NameProceedings - 2012 9th Electronics, Robotics and Automotive Mechanics Conference, CERMA 2012

Conference

Conference2012 9th Electronics, Robotics and Automotive Mechanics Conference, CERMA 2012
Country/TerritoryMexico
CityCuernavaca, Morelos
Period19/11/1223/11/12

Keywords

  • Copyright protection
  • DICOM standard
  • Digital watermarking
  • Geometric distortions
  • Medical imaging

Fingerprint

Dive into the research topics of 'DFT-based watermarking method for medical images'. Together they form a unique fingerprint.

Cite this