A new mammogram image watermarking technique based on two-level intermediate significant bits and sobel edge detector
Recent advances in technology have enabled the digitization of medical images. However, due to the confidentiality requirement for medical images there is the need to ensure that digital medical images are protected from illegal distortions and reproduction. In this research, a new watermarking tech...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
2012
|
| Subjects: | |
| Online Access: | http://eprints.utm.my/id/eprint/41581/1/TinaHajiabdollahMFSKSM2012.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Recent advances in technology have enabled the digitization of medical images. However, due to the confidentiality requirement for medical images there is the need to ensure that digital medical images are protected from illegal distortions and reproduction. In this research, a new watermarking technique based on spatial domain for digital mammogram images is proposed. In the proposed technique, the ROI and RONI of the mammographic image are detected using the Sobel edge detector. Afterwards, the image is partitioned into 16 blocks of the same size, where percentage value of white pixels is calculated for each block. The blocks with the lowest percentage are chosen, which in this case two blocks appear to be of lowest percentage values. The two blocks are of two percent (2%) values each and are used for the embedding process. We then apply BiISB that makes use of two levels of ISB bit-planes (which are the higher-order-ISB and low-order-ISB) at the embedding and extracting phases. The BiISB is employed so as to ensure that the watermarking process is robust and the digital images are made to be imperceptible. For ensuring the security of the digital images, three secret keys were employed at the extraction phase. To verify the robustness and transparency of the proposed technique, NCC and PSNR values are calculated for some sampled images. The results show that the NCC values for the proposed technique maintained high robustness for median filtered images, Salt and Pepper images, Speckle images, Gaussian images, motion effect images, Contrast enhanced images, whereas, the Poisson noise images, rotation and JPEG compressed images maintained values between 0.76 and 1.0. By averaging the PSNR values of all the sampled images, a PSNR value of 59.26 db was achieved. |
|---|