A Discrete Wavelet Based Watermarking Scheme for Authentication of Medical Images

Literature Review

The watermark embedding/extraction are preferably being carried either in frequency or spatial domain. The pixels of a host image or a portion of it are being modified directly by watermark pixels using techniques in spatial domain. e.g. the least significant bit (LSB) (Mehra & Shandilya, 2013). The watermarking techniques in this category exhibit less reliability once exposed to ordinary media operations. Whereas a reversible transformation is applied on the host image using transform domain techniques, before actual watermarking process is being carried out. The transform coefficients so obtained are being used for watermarking purpose (Ghouti et al., 2006; Reddy & Chatterii, 2005; Mehto & Mehra, 2015). Later, inverse transformation is used to obtain the embedded watermark. The techniques in this category however appear to be more robust when exposed to different image related attacks and accordingly provide higher image perceptibility.

Ritu and Manisha (2016) proposed a transform domain based robust watermarking scheme for medical images. The DCT coefficients in mid-band are utilized for embedding the watermark using M-Ary modulation. The patient information also known as electronic patient record is embedded as watermark in the RONI. The higher PSNR, low relative entropy distance as a protection criterion & high robustness are some of the significant features of the proposed scheme.