The field of healthcare is considered to be the most promising application of intelligent sensor networks. However, the security and privacy protection of medical images collected by intelligent sensor networks is a hot problem that has attracted more and more attention. Fortunately, digital watermarking provides an effective method to solve this problem. In order to improve the robustness of the medical image watermarking scheme, in this paper, we propose a novel zero-watermarking algorithm with the integer wavelet transform (IWT), Schur decomposition and image block energy. Specifically, we first use IWT to extract low-frequency information and divide them into non-overlapping blocks, then we decompose the sub-blocks by Schur decomposition. After that, the feature matrix is constructed according to the relationship between the image block energy and the whole image energy. At the same time, we encrypt watermarking with the logistic chaotic position scrambling. Finally, the zero-watermarking is obtained by XOR operation with the encrypted watermarking. Three indexes of peak signal-to-noise ratio, normalization coefficient (NC) and the bit error rate (BER) are used to evaluate the robustness of the algorithm. According to the experimental results, most of the NC values are around 0.9 under various attacks, while the BER values are very close to 0. These experimental results show that the proposed algorithm is more robust than the existing zero-watermarking methods, which indicates it is more suitable for medical image privacy and security protection.
With the increasing maturity of intelligent sensor network technology and the continuous improvement of hospital information construction, intelligent sensor network technology is more and more widely used in hospital information systems by combining wide area networks (WAN), wireless networks and other network fields [
In the current field of healthcare, more and more medical information is collected and transmitted through intelligent sensor networks. Among them, medical images provide a visual way for clinicians to diagnose the condition of patients. An increasing number of medical images are transmitted among different positions through the network. Most of these medical images contain personal privacy details, which may be maliciously intercepted and tampered with by some illegal elements during the transmission. Therefore, the digital watermarking technique has been gradually applied to protect the security of the medical image [
In the spatial domain, the least significant bit (LSB) is a commonly-used technique [
To solve this problem, Wen et al. proposed a zero-watermarking algorithm to construct the watermarking according to the characteristic information of the image itself, which could solve the contradiction between the perceptibility and robustness of digital watermarking [
The current focus of the watermarking algorithm research is on the robustness of this algorithm under various attacks. However, these algorithms are less robust against high-intensity conventional and geometric attacks, especially Gaussian noise, scaling attacks, and cropping attacks. And most algorithms do not test the robustness of multiple attacks.
To address these issues, in this paper, we propose a zero-watermarking algorithm based on IWT, Schur decomposition and image block energy. Specifically, we use IWT to extract low-frequency regions from the original medical image and divide them into non-overlapping blocks, which are subsequently decomposed by the Schur decomposition. Then, we extract the feature matrix by comparing the energy of the image block and the whole image. Finally, we adopt the XOR operation on the encrypted watermarking image and the feature matrix to generate the zero-watermarking. To summarize, we make the following contributions in this work:
We use IWT to avoid the defect of quantization error introduced in the medical image calculation process; We improve the robustness and stability using the Schur decomposition with vector scale invariance and quantum space invariance; We utilize the relationship between the block energy of the transform domain and the average energy of medical images to construct the zero-watermarking, which can achieve good robustness against various attacks even with multiple attacks.
Sweldens et al. proposed a lifting scheme which accelerates the speed of fast wavelet transform [
The IWT is the same as the traditional wavelet transform. The original image is still decomposed into four sub-bands after a wavelet transform. A schematic diagram of the 2-level IWT decomposition is shown in
Schur decomposition is a common matrix decomposition, which is similar to SVD. SVD can be derived from Schur decomposition. Schur decomposition’s theorem is as follows:
For any matrix
The time complexity of SVD is bigger than the time complexity of Schur decomposition [
Each image has its overall energy, but in the zero-watermarking algorithm, the fact that the original image can effectively resist all kinds of attacks shows that the features constructed by the algorithm are very robust. So, on the basis of the overall image energy, choose to block the image to calculate the energy of the block image. The original image is divided into
In [
Attack type | T0 | T1 | PSNR |
---|---|---|---|
None | 2543 | 1553 | — |
Gaussian noise (10%) | 2549 | 1547 | 12.1322 |
JPEG compression (50%) | 2545 | 1551 | 29.5967 |
Median filtering (3 × 3) | 2520 | 1576 | 20.9100 |
Rotation (5°) (anticlockwise) | 2561 | 1535 | 15.5834 |
Scaling (0.25, 4) | 2495 | 1601 | 17.8888 |
As shown in
The medical image watermarking algorithm proposed in this paper can be divided into the process of zero-watermarking construction and extraction, as shown in
The collected medical images are transmitted to the data center by an intelligent sensor network. Then upload the medical image to the medical cloud platform according to the requirements of medical information management. In order to protect patient information disclosure or authentication, the medical image stored in the data center is used as the original image to construct a zero-watermarking.
We use the logistic chaotic scrambling method to encrypt the binary watermarking image
To construct the zero-watermarking, we adopt the IWT to extract the feature matrix from the original image features [
Schur decomposition is a commonly-used matrix decomposition method, which is similar to SVD operation [
Each image has its overall energy, but in the zero-watermarking algorithm, the fact that the original image can effectively resist all kinds of attacks shows that the features constructed by the algorithm are very robust. Because the relationship between the overall average energy of the medical image and the average energy of each block has strong robustness. The calculation of the overall original medical image
On the basis of the overall image energy, we choose to block the image to calculate the energy of the split image. The original image is divided into
Based on the overall average energy and block average energy of the carrier image calculated in the above steps, the method of constructing the feature matrix can be represented as:
Finally, the zero-watermarking
The final zero-watermarking
The zero-watermarking extraction is similar to the above-described construction process. The difference is that the feature matrix obtained from the attacked medical image is XOR with the zero-watermarking from the copyright authentication center, and then inverse scrambling is carried out to the extraction watermarking image
In
In the experiment, we implement the construction and extraction of zero-watermarking using the MATLAB R2018a platform. The original medical images include the brain, lung, chest and hand with the size of 128 × 128 are used as the original carrier images, as shown in
In this paper, we simulate conventional attacks (Gaussian noise, salt & pepper noise, speckle noise, JPEG compression and median filtering, average filtering and Gaussian filtering), geometric attacks (scaling, cropping, rotation, and translation) and combination attacks to evaluate our proposed medical watermarking algorithm. Here, PSNR is used to measure the distortion of the original medical image after being attacked. Besides, we use the NC value given in
Attack type | Attack intensity | Evaluation index | Brain image | Lung image | Chest image | Hand image |
---|---|---|---|---|---|---|
Gaussian noise | 5% | PSNR | 14.7642 | 15.1995 | 14.0572 | 12.3400 |
NC | 1.0000 | 1.0000 | 0.9877 | 0.9627 | ||
BER | 0 | 0 | 0.0156 | 0.0469 | ||
10% | PSNR | 12.0866 | 12.5372 | 11.7086 | 9.9068 | |
NC | 1.0000 | 1.0000 | 0.9502 | 0.9502 | ||
BER | 0 | 0 | 0.0625 | 0.0625 | ||
15% | PSNR | 10.7126 | 10.8986 | 10.5340 | 8.7046 | |
NC | 1.0000 | 0.9877 | 0.9376 | 0.9121 | ||
BER | 0 | 0.0156 | 0.0781 | 0.1094 | ||
20% | PSNR | 9.7727 | 10.1506 | 9.6248 | 7.8319 | |
NC | 0.9877 | 0.9629 | 0.8994 | 0.8993 | ||
BER | 0.0156 | 0.0469 | 0.1248 | 0.1248 | ||
25% | PSNR | 9.1073 | 9.3068 | 9.1064 | 7.3490 | |
NC | 0.9753 | 0.9504 | 0.8865 | 0.8863 | ||
BER | 0.0313 | 0.0625 | 0.1406 | 0.1404 | ||
Salt & pepper noise | 5% | PSNR | 16.7578 | 17.0762 | 18.0804 | 15.8647 |
NC | 1.0000 | 1.0000 | 1.0000 | 0.9877 | ||
BER | 0 | 0 | 0 | 0.0156 | ||
10% | PSNR | 13.6868 | 13.7390 | 14.5451 | 13.0412 | |
NC | 1.0000 | 1.0000 | 0.9877 | 0.9753 | ||
BER | 0 | 0 | 0.0156 | 0.0313 | ||
15% | PSNR | 11.7907 | 12.0737 | 12.8125 | 11.0732 | |
NC | 1.0000 | 1.0000 | 0.9753 | 0.9502 | ||
BER | 0 | 0 | 0.0313 | 0.0625 | ||
20% | PSNR | 10.7667 | 10.8368 | 11.7323 | 10.1200 | |
NC | 1.0000 | 0.9629 | 0.9502 | 0.9376 | ||
BER | 0 | 0.0469 | 0.0625 | 0.0781 | ||
25% | PSNR | 9.7068 | 9.7819 | 10.6847 | 8.9041 | |
NC | 0.9877 | 0.9502 | 0.9249 | 0.9122 | ||
BER | 0.0156 | 0.0625 | 0.0938 | 0.1091 | ||
Speckle noise | 5% | PSNR | 21.1982 | 21.4382 | 17.7929 | 22.2242 |
NC | 1.0000 | 1.0000 | 1.0000 | 1.0000 | ||
BER | 0 | 0 | 0 | 0 | ||
10% | PSNR | 18.2521 | 19.0031 | 15.1206 | 19.1629 | |
NC | 1.0000 | 1.0000 | 1.0000 | 1.0000 | ||
BER | 0 | 0 | 0 | 0 | ||
15% | PSNR | 16.6040 | 17.4271 | 13.5238 | 17.5106 | |
NC | 1.0000 | 1.0000 | 0.9877 | 0.9877 | ||
BER | 0 | 0 | 0.0156 | 0.0156 | ||
20% | PSNR | 15.5492 | 16.3898 | 12.4485 | 16.2019 | |
NC | 1.0000 | 1.0000 | 0.9877 | 0.9877 | ||
BER | 0 | 0 | 0.0156 | 0.0156 | ||
25% | PSNR | 14.7916 | 15.6281 | 11.5182 | 15.3433 | |
NC | 1.0000 | 1.0000 | 0.9502 | 0.9877 | ||
BER | 0 | 0 | 0.0625 | 0.0156 | ||
JPEG compression | 5% | PSNR | 22.0045 | 23.9372 | 26.7095 | 26.8927 |
NC | 1.0000 | 1.0000 | 0.9753 | 0.9877 | ||
BER | 0 | 0 | 0.0315 | 0.0153 | ||
10% | PSNR | 23.8912 | 26.9916 | 29.8869 | 29.6461 | |
NC | 1.0000 | 1.0000 | 0.9877 | 1.0000 | ||
BER | 0 | 0 | 0.0156 | 0 | ||
15% | PSNR | 24.9075 | 28.465 | 31.7934 | 31.1542 | |
NC | 1.0000 | 1.0000 | 0.9877 | 0.9877 | ||
BER | 0 | 0 | 0.0156 | 0.0156 | ||
20% | PSNR | 25.8223 | 29.5188 | 32.9327 | 32.1163 | |
NC | 1.0000 | 1.0000 | 0.9877 | 0.9877 | ||
BER | 0 | 0 | 0.0156 | 0.0156 | ||
25% | PSNR | 26.7087 | 30.3743 | 33.8355 | 32.8303 | |
NC | 1.0000 | 1.0000 | 0.9877 | 1.0000 | ||
BER | 0 | 0 | 0.0156 | 0 | ||
Median filtering | 3 × 3 | PSNR | 23.5681 | 29.2442 | 37.9383 | 35.8022 |
NC | 1.0000 | 1.0000 | 1.0000 | 1.0000 | ||
BER | 0 | 0 | 0 | 0 | ||
5 × 5 | PSNR | 19.4000 | 23.9270 | 32.2365 | 31.9462 | |
NC | 1.0000 | 1.0000 | 0.9877 | 1.0000 | ||
BER | 0 | 0 | 0.0156 | 0 | ||
7 × 7 | PSNR | 17.3484 | 22.2896 | 28.8159 | 28.9450 | |
NC | 1.0000 | 1.0000 | 0.9753 | 1.0000 | ||
BER | 0 | 0 | 0.0313 | 0 | ||
9 × 9 | PSNR | 15.8117 | 21.3870 | 26.6631 | 25.8972 | |
NC | 1.0000 | 1.0000 | 0.9753 | 0.9627 | ||
BER | 0 | 0 | 0.0313 | 0.0469 | ||
11 × 11 | PSNR | 14.8774 | 20.5817 | 25.1021 | 23.4141 | |
NC | 1.0000 | 1.0000 | 0.9628 | 0.9627 | ||
BER | 0 | 0 | 0.0469 | 0.0469 | ||
Average filtering | 3 × 3 | PSNR | 21.3212 | 26.6291 | 29.4174 | 30.8675 |
NC | 1.0000 | 1.0000 | 1.0000 | 1.0000 | ||
BER | 0 | 0 | 0 | 0 | ||
5 × 5 | PSNR | 18.4333 | 22.3107 | 25.9712 | 27.4251 | |
NC | 1.0000 | 1.0000 | 0.9877 | 1.0000 | ||
BER | 0 | 0 | 0.0156 | 0 | ||
7 × 7 | PSNR | 17.0003 | 20.2281 | 23.8956 | 25.2664 | |
NC | 1.0000 | 1.0000 | 0.987 | 0.9877 | ||
BER | 0 | 0 | 0.0156 | 0.0156 | ||
9 × 9 | PSNR | 16.1408 | 18.8448 | 22.4341 | 23.4856 | |
NC | 1.0000 | 1.0000 | 0.9877 | 0.9753 | ||
BER | 0 | 0 | 0.0156 | 0.0313 | ||
11 × 11 | PSNR | 15.6052 | 17.7630 | 21.3092 | 22.0736 | |
NC | 1.0000 | 0.9879 | 0.9877 | 0.9753 | ||
BER | 0 | 0.0154 | 0.0156 | 0.0313 | ||
Gaussian filtering | 3 × 3 | PSNR | 21.4444 | 26.7688 | 29.5766 | 31.0254 |
NC | 1.0000 | 1.0000 | 1.0000 | 1.0000 | ||
BER | 0 | 0 | 0 | 0 | ||
5 × 3 | PSNR | 18.7563 | 22.7303 | 26.3948 | 27.8346 | |
NC | 1.0000 | 1.0000 | 0.9877 | 1.0000 | ||
BER | 0 | 0 | 0.0156 | 0 | ||
7 × 3 | PSNR | 17.5649 | 20.9820 | 24.7006 | 26.0922 | |
NC | 1.0000 | 1.0000 | 0.9877 | 1.0000 | ||
BER | 0 | 0 | 0.0156 | 0 | ||
9 × 3 | PSNR | 16.9531 | 20.0162 | 23.7148 | 24.9039 | |
NC | 1.0000 | 1.0000 | 0.9877 | 0.9753 | ||
BER | 0 | 0 | 0.0156 | 0.0313 | ||
11 × 3 | PSNR | 16.6365 | 19.4445 | 23.1469 | 24.1661 | |
NC | 1.0000 | 1.0000 | 0.9877 | 0.9753 | ||
BER | 0 | 0 | 0.0156 | 0.0313 |
Attack type | Attack intensity | Evaluation index | Brain image | Lung image | Chest image | Hand image |
---|---|---|---|---|---|---|
Scaling | 0.125, 8 | PSNR | 15.6243 | 17.6118 | 24.5495 | 22.0149 |
NC | 1.0000 | 1.0000 | 0.9877 | 0.9627 | ||
BER | 0 | 0 | 0.0157 | 0.0469 | ||
0.25, 4 | PSNR | 17.8888 | 21.8774 | 29.2038 | 28.1517 | |
NC | 1.0000 | 1.0000 | 0.9877 | 1.0000 | ||
BER | 0 | 0 | 0.0157 | 0 | ||
0.5, 2 | PSNR | 21.5809 | 27.0624 | 35.8517 | 32.9718 | |
NC | 1.0000 | 1.0000 | 1.0000 | 1.0000 | ||
BER | 0 | 0 | 0 | 0 | ||
2, 0.5 | PSNR | 30.2404 | 38.8692 | 47.1327 | 43.5692 | |
NC | 1.0000 | 1.0000 | 1.0000 | 1.0000 | ||
BER | 0 | 0 | 0 | 0 | ||
4, 0.25 | PSNR | 30.5191 | 39.1329 | 47.5906 | 43.8283 | |
NC | 1.0000 | 1.0000 | 1.0000 | 1.0000 | ||
BER | 0 | 0 | 0 | 0 | ||
5% | PSNR | 90.1377 | 22.2412 | 19.0604 | 25.2433 | |
NC | 1.0000 | 1.0000 | 0.9628 | 1.0000 | ||
BER | 0 | 0 | 0.0469 | 0 | ||
10% | PSNR | 90.1377 | 18.4078 | 14.7486 | 22.0753 | |
NC | 1.0000 | 0.9752 | 0.9376 | 0.98766 | ||
BER | 0 | 0.0313 | 0.0781 | 0.0156 | ||
Cropping | 15% | PSNR | 56.9175 | 16.1173 | 12.4962 | 20.5341 |
NC | 1.0000 | 0.9752 | 0.8865 | 0.9877 | ||
BER | 0 | 0.0313 | 0.1406 | 0.1563 | ||
20% | PSNR | 25.9643 | 14.8096 | 11.1329 | 19.3096 | |
NC | 1.0000 | 0.9376 | 0.8603 | 0.9753 | ||
BER | 0 | 0.0779 | 0.1719 | 0.0312 | ||
25% | PSNR | 19.9237 | 13.9273 | 10.3298 | 17.7532 | |
NC | 1.0000 | 0.92476 | 0.8603 | 0.9753 | ||
BER | 0 | 0.0935 | 0.1719 | 0.0313 | ||
Rotation | 3° | PSNR | 18.2249 | 19.3431 | 20.3276 | 21.9927 |
NC | 1.0000 | 1.0000 | 0.9753 | 1.0000 | ||
BER | 0 | 0 | 0.0313 | 0 | ||
5° | PSNR | 15.5834 | 16.6031 | 17.3382 | 18.9684 | |
NC | 1.0000 | 0.9877 | 0.9501 | 0.9627 | ||
BER | 0 | 0.0156 | 0.0625 | 0.0469 | ||
10° | PSNR | 12.9947 | 13.2794 | 13.7619 | 15.9034 | |
NC | 1.0000 | 0.9753 | 0.8994 | 0.8993 | ||
BER | 0 | 0.0313 | 0.1248 | 0.1250 | ||
15° | PSNR | 12.2902 | 11.4540 | 11.9394 | 14.9546 | |
NC | 0.9501 | 0.9252 | 0.7984 | 0.8863 | ||
BER | 0.0625 | 0.0938 | 0.2498 | 0.1406 | ||
20° | PSNR | 12.0683 | 10.1806 | 10.8055 | 14.3554 | |
NC | 0.9376 | 0.8347 | 0.7277 | 0.8465 | ||
BER | 0.0781 | 0.2031 | 0.3279 | 0.1873 | ||
Translation | 1% | PSNR | 17.8600 | 20.2850 | 25.4793 | 30.1726 |
NC | 1.0000 | 0.9877 | 0.9877 | 1.0000 | ||
BER | 0 | 0.0156 | 0.0156 | 0 | ||
3% | PSNR | 13.2478 | 14.1042 | 19.4053 | 24.0885 | |
NC | 0.9753 | 0.9877 | 0.9627 | 1.0000 | ||
BER | 0.0313 | 0.0156 | 0.0469 | 0 | ||
5% | PSNR | 11.9854 | 11.0666 | 15.9737 | 20.6979 | |
NC | 0.9503 | 0.8473 | 0.9251 | 1.0000 | ||
BER | 0.0625 | 0.1875 | 0.0938 | 0 | ||
7% | PSNR | 11.8748 | 9.8692 | 14.6068 | 19.2970 | |
NC | 0.9503 | 0.7799 | 0.9125 | 0.9877 | ||
BER | 0.0625 | 0.2656 | 0.1094 | 0.0156 | ||
9% | PSNR | 11.6587 | 8.8945 | 13.0342 | 17.7887 | |
NC | 0.9247 | 0.7249 | 0.8732 | 0.9376 | ||
BER | 0.0938 | 0.3281 | 0.1563 | 0.0781 |
Attack type | Attack intensity | Evaluation index | Brain image | Lung image | Chest image | Hand image |
---|---|---|---|---|---|---|
Gaussian noise and median filtering | 5%, 3 × 3 | PSNR | 19.1489 | 20.9061 | 20.7881 | 18.8019 |
NC | 1.0000 | 1.0000 | 1.0000 | 1.0000 | ||
BER | 0 | 0 | 0 | 0 | ||
10%, 5 × 5 | PSNR | 17.5993 | 20.1135 | 20.6786 | 19.1843 | |
NC | 1.0000 | 1.0000 | 1.0000 | 1.0000 | ||
BER | 0 | 0 | 0 | 0 | ||
15%, 7 × 7 | PSNR | 16.3756 | 19.0191 | 20.0520 | 18.9095 | |
NC | 1.0000 | 1.0000 | 0.9877 | 1.0000 | ||
BER | 0 | 0 | 0.0156 | 0 | ||
20%, 9 × 9 | PSNR | 15.6893 | 18.1736 | 18.7612 | 18.4873 | |
NC | 1.0000 | 1.0000 | 0.9877 | 0.9877 | ||
BER | 0 | 0 | 0.0156 | 0.0156 | ||
25%, 11 × 11 | PSNR | 15.2340 | 17.1140 | 17.8346 | 17.9449 | |
NC | 1.0000 | 1.0000 | 0.9753 | 0.9877 | ||
BER | 0 | 0 | 0.0313 | 0.0156 | ||
JPEG compression and median filtering | 5%, 3 × 3 | PSNR | 21.1295 | 24.5684 | 27.3012 | 27.5828 |
NC | 1.0000 | 1.0000 | 0.9753 | 0.9877 | ||
BER | 0 | 0 | 0.0313 | 0.0156 | ||
10%, 5 × 5 | PSNR | 19.2976 | 24.1588 | 29.7951 | 29.1742 | |
NC | 1.0000 | 1.0000 | 0.9877 | 1.0000 | ||
BER | 0 | 0 | 0.0156 | 0 | ||
15%, 7 × 7 | PSNR | 17.4761 | 22.3819 | 28.3498 | 27.9810 | |
NC | 1.0000 | 1.0000 | 0.9753 | 0.9877 | ||
BER | 0 | 0 | 0.0313 | 0.0156 | ||
20%, 9 × 9 | PSNR | 16.0470 | 21.4425 | 26.4696 | 25.7769 | |
NC | 1.0000 | 1.0000 | 0.9753 | 0.9627 | ||
BER | 0 | 0 | 0.0313 | 0.0469 | ||
25%, 11 × 11 | PSNR | 15.1197 | 20.6040 | 25.0580 | 23.6894 | |
NC | 1.0000 | 1.0000 | 0.9628 | 0.9627 | ||
BER | 0 | 0 | 0.0469 | 0.0469 | ||
Cropping and median filtering | 5%, 3 × 3 | PSNR | 23.5681 | 21.4882 | 19.0059 | 24.8774 |
NC | 1.0000 | 1.0000 | 0.9628 | 1.0000 | ||
BER | 0 | 0 | 0.0469 | 0 | ||
10%, 5 × 5 | PSNR | 19.4000 | 17.4241 | 14.6643 | 21.6490 | |
NC | 1.0000 | 0.9752 | 0.9376 | 0.9877 | ||
BER | 0 | 0.0313 | 0.0781 | 0.0156 | ||
15%, 7 × 7 | PSNR | 17.3443 | 15.2166 | 12.3925 | 19.9473 | |
NC | 1.0000 | 0.9503 | 0.8865 | 0.9752 | ||
BER | 0 | 0.0623 | 0.1406 | 0.0313 | ||
20%, 9 × 9 | PSNR | 15.5347 | 13.9957 | 11.0146 | 18.4451 | |
NC | 1.0000 | 0.9376 | 0.8603 | 0.9753 | ||
BER | 0 | 0.0779 | 0.1719 | 0.0313 | ||
25%, 11 × 11 | PSNR | 14.1811 | 13.1628 | 10.1960 | 16.7092 | |
NC | 1.0000 | 0.9248 | 0.8735 | 0.9877 | ||
BER | 0 | 0.0935 | 0.1563 | 0.0156 | ||
JPEG compression and scaling | 5%, 0.125 | PSNR | 15.5163 | 17.4607 | 24.2929 | 21.7789 |
NC | 1.0000 | 1.0000 | 0.9753 | 0.9627 | ||
BER | 0 | 0 | 0.0313 | 0.0469 | ||
10%, 0.25 | PSNR | 17.7903 | 21.7549 | 28.5464 | 27.5526 | |
NC | 1.0000 | 1.0000 | 0.9877 | 1.0000 | ||
BER | 0 | 0 | 0.0156 | 0 | ||
15%, 0.5 | PSNR | 21.2622 | 26.3763 | 32.6142 | 31.2318 | |
NC | 1.0000 | 1.0000 | 0.9877 | 0.9877 | ||
BER | 0 | 0 | 0.0156 | 0.0156 | ||
20%, 2.0 | PSNR | 25.5233 | 30.0743 | 33.4944 | 32.5491 | |
NC | 1.0000 | 1.0000 | 0.9877 | 0.9877 | ||
BER | 0 | 0 | 0.0156 | 0.0156 | ||
25%, 4.0 | PSNR | 26.2767 | 30.8143 | 34.4413 | 33.3297 | |
NC | 1.0000 | 1.0000 | 0.9877 | 1.0000 | ||
BER | 0 | 0 | 0.0156 | 0 | ||
0.125, 5% | PSNR | 15.7607 | 16.4132 | 18.1330 | 22.2645 | |
NC | 1.0000 | 1.0000 | 0.9628 | 0.9753 | ||
BER | 0 | 0 | 0.0469 | 0.0313 | ||
0.25, 10% | PSNR | 18.0261 | 16.8881 | 14.7144 | 23.0562 | |
NC | 1.0000 | 0.9752 | 0.9376 | 0.9877 | ||
BER | 0 | 0.0313 | 0.0781 | 0.0156 | ||
Scaling and cropping | 0.5, 15% | PSNR | 21.7197 | 15.8221 | 12.5822 | 22.2321 |
NC | 1.0000 | 0.9629 | 0.8865 | 0.9877 | ||
BER | 0 | 0.0466 | 0.1406 | 0.0156 | ||
2, 20% | PSNR | 24.7837 | 14.7955 | 11.2348 | 21.2317 | |
NC | 1.0000 | 0.9376 | 0.8603 | 0.9753 | ||
BER | 0 | 0.0779 | 0.1719 | 0.0313 | ||
4, 25% | PSNR | 19.7524 | 13.9171 | 10.4319 | 19.6812 | |
NC | 1.0000 | 0.9248 | 0.8603 | 0.9753 | ||
BER | 0 | 0.0935 | 0.1719 | 0.0313 | ||
Rotation and scaling | 3°, 0.125 | PSNR | 15.7016 | 17.0976 | 21.9632 | 22.9540 |
NC | 1.0000 | 1.0000 | 0.9753 | 0.9753 | ||
BER | 0 | 0 | 0.0313 | 0.0313 | ||
5°, 0.25 | PSNR | 16.7363 | 17.5117 | 18.5239 | 21.7308 | |
NC | 1.0000 | 0.9877 | 0.9501 | 0.9627 | ||
BER | 0 | 0.0156 | 0.0625 | 0.0469 | ||
10°, 0.5 | PSNR | 13.8861 | 13.5584 | 14.0713 | 18.0076 | |
NC | 0.9877 | 0.9753 | 0.8994 | 0.8993 | ||
BER | 0.0156 | 0.0313 | 0.1248 | 0.1250 | ||
15°, 2.0 | PSNR | 12.5631 | 11.4826 | 12.0792 | 16.9252 | |
NC | 0.9501 | 0.9378 | 0.7954 | 0.8863 | ||
BER | 0.0625 | 0.0781 | 0.2498 | 0.1406 | ||
20°, 4.0 | PSNR | 12.3288 | 10.2007 | 10.9375 | 16.3234 | |
NC | 0.9376 | 0.8347 | 0.7277 | 0.8465 | ||
BER | 0.0781 | 0.2031 | 0.3279 | 0.1873 |
The first is to carry out some kinds of noise attacks. In this paper, Gaussian noise, salt & pepper noise and speckle noise with different intensities (5%, 10%, 15%, 20% and 25%) are selected to attack the four original medical images. The NC values of all four medical images under the Gaussian noise attack are above 0.88 with the increasing noise intensity as shown in
Similarly, the NC values of all four medical images are maintained above 0.97 and their BER values are less than 0.04 through JPEG compression attacks of different intensities. When the strength is 15%, the attack results of the four images are shown in
Three kinds of filtering attacks are also selected in this paper, namely, median filtering, average filtering, and Gaussian filtering. As can be seen from
For medical images such as the chest radiograph, when subjected to a slightly stronger cropping attack, the pixels of the extracted watermarking image are somewhat distorted, but the average value of NC is still higher than 0.86.
Image rotation is a common geometric attack that changes the position of image pixels. After rotating 10 degrees counterclockwise, as shown in
In the process of downward translation, the PSNR obtained from the original medical image decreases gradually, but the NC values are high enough and the BER values are also close to 0. In
Six combined attacks are selected in this paper. The first is the combination of two conventional attacks. The first kind is the Gaussian noise attack on the medical image and then the median filtering attack, and the second is the JPEG compression attack on the medical image and then the median filtering attack. As can be seen in
The second major category is the combination of conventional attacks and geometric attacks. The first one cuts the image along the X-axis and then performs the median filtering attack. Compared with the four images, the NC values of the chest image are slightly lower but still above 0.85, and the NC values of the other three images remain above 0.9. The second attack is the JPEG compression attack on the image and then the scaling attack. In
The third category of attacks is two geometric attacks on images, namely, scaling attack combined with cropping attack, and rotation attack combined with scaling attack. From the data in
To verify the advantage of our proposed algorithm, we used the same experiment condition to compare it with other representative works [
Attack type | Attack intensity | Proposed algorithm | Algorithm [ |
Algorithm [ |
Algorithm [ |
Algorithm [ |
---|---|---|---|---|---|---|
Gaussian noise | 5% | 1.0000 | 0.9375 | 0.9749 | 0.6830 | 0.7366 |
10% | 1.0000 | 0.9504 | 0.9514 | 0.6667 | 0.6671 | |
15% | 1.0000 | 0.9504 | 0.9199 | 0.6555 | 0.6447 | |
JPEG compression | 5% | 1.0000 | 0.9877 | 0.9863 | 0.5748 | 0.6738 |
10% | 1.0000 | 0.9502 | 0.9882 | 0.6518 | 0.9863 | |
15% | 1.0000 | 0.9629 | 0.9935 | 0.7126 | 0.8530 | |
Median filtering | 3 × 3 | 1.0000 | 0.9377 | 0.9888 | 0.7088 | 0.9969 |
5 × 5 | 1.0000 | 0.9123 | 0.9713 | 0.7511 | 0.9649 | |
7 × 7 | 1.0000 | 0.9123 | 0.9532 | 0.7671 | 0.9579 |
Attack type | Attack intensity | Proposed algorithm | Algorithm [ |
Algorithm [ |
Algorithm [ |
Algorithm [ |
---|---|---|---|---|---|---|
Gaussian noise | 5% | 0 | 0.0781 | 0.0332 | 0.5313 | 0.3127 |
10% | 0 | 0.0625 | 0.0637 | 0.5469 | 0.4143 | |
15% | 0 | 0.0625 | 0.1028 | 0.6250 | 0.4263 | |
JPEG compression | 5% | 0 | 0.0156 | 0.0173 | 0.4844 | 0.3831 |
10% | 0 | 0.0625 | 0.0149 | 0.4063 | 0.0173 | |
15% | 0 | 0.0469 | 0.0083 | 0.3438 | 0.1731 | |
Median filtering | 3 × 3 | 0 | 0.0781 | 0.0142 | 0.3438 | 0.0039 |
5 × 5 | 0 | 0.1094 | 0.0361 | 0.2966 | 0.0442 | |
7 × 7 | 0 | 0.1094 | 0.0586 | 0.2810 | 0.0505 |
Attack type | Attack intensity | Proposed algorithm | Algorithm |
Algorithm [ |
Algorithm [ |
Algorithm [ |
---|---|---|---|---|---|---|
Scaling | 0.125, 8 | 1.0000 | 0.8746 | 0.9284 | 0.6690 | 0.8037 |
0.25, 4 | 1.0000 | 0.9628 | 0.9670 | 0.6989 | 0.8763 | |
0.5, 2 | 1.0000 | 0.9753 | 0.9908 | 0.7372 | 0.9579 | |
Cropping | 20% | 1.0000 | 0.9753 | 0.9969 | 0.9754 | 0.9776 |
25% | 1.0000 | 0.9377 | 0.9834 | 0.9377 | 0.9316 | |
30% | 1.0000 | 0.9502 | 0.9655 | 0.8998 | 0.9214 | |
Rotation | 5° | 1.0000 | 0.9628 | 0.9441 | 0.6649 | 0.9492 |
10° | 1.0000 | 0.9753 | 0.9089 | 0.6351 | 0.8890 | |
15° | 0.9501 | 0.9502 | 0.8901 | 0.6666 | 0.8619 | |
Translation | 5% | 0.9503 | 0.9378 | 0.8996 | 0.7532 | 0.8840 |
10% | 0.9247 | 0.9378 | 0.8819 | 0.7106 | 0.8371 | |
15% | 0.9118 | 0.9753 | 0.8548 | 0.6966 | 0.7920 |
Attack type | Attack intensity | Proposed algorithm | Algorithm [ |
Algorithm [ |
Algorithm [ |
Algorithm [ |
---|---|---|---|---|---|---|
Scaling | 0.125, 8 | 0 | 0.1563 | 0.0891 | 0.3906 | 0.2388 |
0.25, 4 | 0 | 0.0469 | 0.0415 | 0.3591 | 0.1526 | |
0.5, 2 | 0 | 0.0313 | 0.0117 | 0.3125 | 0.0530 | |
Cropping | 20% | 0 | 0.0313 | 0.0039 | 0.0313 | 0.0283 |
25% | 0 | 0.0781 | 0.0210 | 0.0781 | 0.0857 | |
30% | 0 | 0.0625 | 0.0435 | 0.1248 | 0.0984 | |
Rotation | 5° | 0 | 0.0469 | 0.0698 | 0.3904 | 0.0637 |
10° | 0 | 0.0313 | 0.1133 | 0.4216 | 0.1375 | |
15° | 0.0625 | 0.0625 | 0.1355 | 0.3904 | 0.1694 | |
Translation | 5% | 0.0625 | 0.0781 | 0.1235 | 0.2966 | 0.1431 |
10% | 0.0938 | 0.0781 | 0.1450 | 0.3435 | 0.1985 | |
15% | 0.1094 | 0.0313 | 0.1768 | 0.3591 | 0.2498 |
Proposed algorithm | Algorithm [ |
Algorithm [ |
Algorithm [ |
Algorithm [ |
|
---|---|---|---|---|---|
Average times (sec) | 0.9406 | 1.0975 | 0.9524 | 0.9540 | 1.3329 |
When carrying out conventional attacks, as shown in
As shown on the right in
To more comprehensively detect the algorithm’s performance, under the same experimental environment, the computing time of the proposed algorithm and the literature [
Aiming to protect the security of the medical image and not damage original information, we have proposed a new zero-watermarking algorithm in this work. Specifically, we used the IWT to extract low-frequency information from the original medical image, which was then divided into blocks by the Schur decomposition. After that, we constructed the feature matrix according to the relation between image block energy. Meanwhile, we encrypted the watermarking information using logistic position scrambling. Finally, zero-watermarking is generated via the XOR operation between the scrambled watermarking information and the feature matrix. We compared our algorithm with other representative works under a series of conventional attacks and geometric attacks in the experiment. Experimental results show that the proposed algorithm could improve the robustness of the medical image zero-watermarking, especially for the high-intensity of conventional attacks and geometric attacks. This algorithm can efficiently ensure the safety and privacy of patients and the confidentiality and reliability of medical images. However, this algorithm in this paper is aimed at 2D medical images. It has not been applied to 3D medical images, our future work will consider applying the proposed algorithm to the protection of 3D medical images, and we will attempt to design a robust zero-watermarking algorithm that can protect 2D and 3D medical images.