In this section, the proposed methodology is evaluated using 100MRI cover images (all of them square measure gray level of size 255 × 255). in the experiments, stego images are created using the proposed technique with different embedding rates starting from five-hitter to five hundredth. Then image features are extracted from the cover and stego pictures.
the model has been developed using MATLAB 2017a software system tool, wherever a well-developed script file has been developed to feed information|the info|the information} (medical image or the cover data and patient information). The developed script facilitates user to pick input file, and user outlined key series to perform encrypt secret data at the transmitter end, before embedding it to the cover image. further at the receiver to perform extraction, coding secret’s used to retrieve the key key; but before it STC and hamming code applied to decompose stego image and retrieve hidden information from coefficients. Quality Measurement:-The Quality of the reconstructed image is measured in-terms of mean square error, peak signal to noise ratio, embedding capacity, weighted Peak signal-to-noise ratio and average difference. Mean square errorThe MSE is commonly known as reconstruction error variance ?q2. The MSE between the first image f and the reconstructed image g at decoder is outlined as: MSE = Where the add over j, k denotes the add over all pixels within the image and N is that the range of pixels in every image. Peak signal-to-noise ratioFrom that the peak {signal-to-noise ratio|signal-to-noise|signal/noise ratio|signal/noise|S/N|ratio} is outlined because the ratio between signal variance and reconstruction error variance.
The PSNR between 2 images having eight bits per pixel in terms of decibels (dB) is given by:PSNR = 10 log10 Generally once PSNR is 20 db or bigger, then the first and the reconstructed images square measure nearly in-distinguishable by human eyes.Embedding capacityTo evaluate the performance of the proposed scheme, the length of the key message (data capacity) is employed in concert of the analysis criteria, that is outlined because the quantity of bits which will be embedded into the cover image. The embedding capacity is given by:E = K/WH (bpp)Where K is that the range of information message bits, whereas W and Hare the dimension and height of {the cover|the duvet|the quilt} image severally (both cover and stego pictures square measure of identical size with W = H = 255).Weighted Peak signal-to-noise ratioThe weighted Peak signal-to-noise ratio (wPSNR) is associate alternate activity of physical property. It utilizes an additional parameter known as Noise Visibility function (NVF). wPSNR is roughly such as PSNR for flat areas as a result of NVF is closeto one in sleek regions. However, for regions with sharp contrasts, wPSNR is on top of PSNR, as a result of NVF is near zero for advanced regions. Hence, wPSNR makes an attempt to replicate however the Human visual system (HVS) perceives images.
wPSNR = 10 log10 Average differenceThe average difference may be a easy and popular image quality analysis criterion. it’s computed by averaging absolutely the distinction between the cover and stego images, that is calculated as shown inAverage distinction = The stego images created with embedding rates from five-hitter to four-hundredth using the proposed technique using STC and hamming code are shown in Figure 8 and figure 9. One will notice that it’s difficult for the human eye to differentiate between the first and stego pictures. Moreover, the histogram of the stego pictures (represented in Figure 8 and figure 9) square measure quite similar to that of the cover images. it is it is necessary to say that for a 255 × 255 image, that contains 65,025 pixels, associate embedding rate of 400th means embedding (40/100) * 65,025 = 26,010 bits, which is equivalent to 3715 ascii characters. For a 512 × 512 image, associate embed-ding rate of 400th means embedding 104,757 bits or 14,979 ascii characters. Figure 8: STC Stego Image and histogram of Stego Image.
Figure 9: hamming code Stego Image and histogram of Stego Image.In order to possess a comprehensive comparison with the proposed strategies, we tend to implemented LSB-RSA and LSB-AES methods are shown in figure 10 and figure 11. In LSB-RSA, we tend to introduce the RSA encrypted secret knowledge in the in the medical image by using the smallest amount important bit. In LSB_AES technique, we tend to expand LSB steganography technique and used our edge detection technique to find the sharpest regions for the embedding method.
And introduce the AES encrypted secret knowledge into image by victimisation the smallest amount important bit. Figure 10: RSA Stego Image and histogram of Stego Image. Figure 11: AES Stego Image and histogram of Stego Image.The visual quality performance results are shown in Table 1 to Table 4.
it’s clear that the proposed technique using STC and hamming code, rather than STC obtained the best image quality in all image metric measurements compared to the other ways, followed by the hamming code implementation of the proposed algorithmic rule, LSB-AES and LSB-RSA.Table 1: Comparison of the results of hamming Code technique. Table 2: Comparison of the results of STC technique. Table 3: Comparison of the results of RSA technique. Table 4: Comparison of the results of AES technique.
A graphical illustration of the PSNR and wPSNR values are shown in Figure 12 and Figure 13 severally. Those 2 figures demonstrate the superior performance of our proposed technique using STC, and to a slightly less degree using hamming code, compared to the other methods like RSA and AES, with PSNR values greater than 50 db and wPSNR values greater than 55 db, which suggests the hidden data is undetectable in step with the human visual perception. Figure 12:- Performance analysis for PSNR values for STC, hamming Code, RSA and AES using different embedding rates. Figure 13:- Performance analysis for WPSNR values for STC, hamming Code, RSA and AES using different embedding rates.
