Digital Watermarking of Medical Image Fusion based Discrete Wavelet Transform and Singular Value Decomposition

碩士 === 國立臺灣海洋大學 === 通訊與導航工程學系 === 103 === Now a days many hospitals and diagnostic centers have started using wireless media for transmitting and receiving Medical information, security for exchanging these Medical information is highly required. In our study we propose Digital Watermarking of Fused...

Full description

Bibliographic Details
Main Authors: Farau W. Sako, 沙威廉
Other Authors: Jia-Chyi, Wu
Format: Others
Language:en_US
Published: 2015
Online Access:http://ndltd.ncl.edu.tw/handle/39862109168805695420
Description
Summary:碩士 === 國立臺灣海洋大學 === 通訊與導航工程學系 === 103 === Now a days many hospitals and diagnostic centers have started using wireless media for transmitting and receiving Medical information, security for exchanging these Medical information is highly required. In our study we propose Digital Watermarking of Fused Medical Image based on Discrete wavelet transform and Singular value decomposition algorithm, by applying digital watermarking to the medical image not only protecting medical image from the third part (attackers) but also providing authentication by means of adding ownership data as the watermark image in the medical information. We apply Discrete wavelet transform and Singular value decomposition algorithm in our study, we first fuse the two source images from two modalities of MRI and CT imaging, we apply the first level of decomposition of DWT to the Medical fused image to obtain low and high frequency coefficients bands, we then apply SVD to each of the frequency sub-band of the medical image. We decompose also the watermark image into low and frequency coefficient bands, and then to the low frequency coefficient bands of watermark image we apply the SVD. Then we perform digital watermarking by embedding low frequency coefficient of watermark image to the modified frequency coefficient of Medical Fused Image. We have also examine our algorithm against number of attacks, simulations show that the algorithm is more robust against noise attacks.