A Highly Robust Medical Image Watermarking Method for Medical Real-time Applications
Mahdi Mehrabi1, Vahdi Zarei1, Mohammad Ghanbari2
1Department of Computer Engineering, Shiraz Branch, Islamic Azad University, Shiraz, Iran.
Background:
Watermarking such as other security concepts is an ongoing challenging research issue, especially for medical images, to protect patient privacy. Medical images need to be shared and transferred between hospitals and specialists as quickly as possible for better diagnosis. Fast and simple watermarking is needed as well as the robust transferring of channel noise, such as salt and pepper noise and robust cropping that may occur from specialists and signature encryption for patient privacy.
Methods:
In this article, a highly robust and simple watermarking method is introduced. The proposed method has very low computational complexity and at the same time, it is very robust to interference and uses simple computations such as (XORs) Exclusive ORs and rotations that can be done in real-time. The proposed method uses a combination of hidden neighboring signature information, Sudoku permutation, and noise pre-processing to achieve high robustness against salt and pepper noise and cropping. Simple signature encryption is also used.
Results:
The proposed method is examined in different medical image datasets. The experimental results indicate the proposed watermarking system is robust to salt and pepper noise density of up to 90% and about 70% cropping. The number of computations including encryption is five XOR per pixel and a rotation per block of signature size.
Conclusion:
A novel method for medical image watermarking is presented. The proposed method is in the spatial domain, has encryption, and uses only XOR computation. The proposed method is highly robust to noise and cropping which is necessary for medical uses. The proposed method can be used efficiently for real-time watermarking for medical and nonmedical image datasets.
