Double Watermarking Algorithm for Tamper Detection of Medical Images

Miaomiao Wang; Deyang Wu; Sen Hu; Jiayan Wang; Yan Wang; Haibo Jin; Changbo Qu

doi:10.3788/LOP213052

Laser & Optoelectronics Progress, Volume. 60, Issue 6, 0617001(2023)

Double Watermarking Algorithm for Tamper Detection of Medical Images

Miaomiao Wang1... Deyang Wu2,3,*, Sen Hu1, Jiayan Wang2,3, Yan Wang1, Haibo Jin1 and Changbo Qu1 |Show fewer author(s)

Author Affiliations

¹College of Software, Liaoning Technical University, Huludao 125105, Liaoning, China

²College of Information Science and Engineering, Yan Shan University, Qinhuangdao 066004, Hebei, China

³The Key Laboratory for Computer Virtual Technology and System Integration of Hebei Province, Qinhuangdao 066004, Hebei, China

show less

Abstract Get PDF(in Chinese)

Figures & Tables(14)

Fig. 1. Construction principle and effect of safety area. (a) Original image; (b) safe area construction; (c) rotation 45°; (d) rotation correction

Download full size

Fig. 2. Chaotic time domain diagram

Download full size

Fig. 3. Double watermark construction process

Download full size

Fig. 4. Encryption process of watermark image

Download full size

Fig. 5. Double watermark extraction process

Download full size

Fig. 6. Host image and logo image. (a) Fundus 1; (b) fundus 2; (c) fundus 3; (d) fundus 4; (e) brain map 1; (f) brain map 2; (g) brain map 3; (h) brain map 4; (i) copyright logo

Download full size

Fig. 7. Tamper detection. (a) No attack; (b) Gaussian noise 0.005; (c) median filter 11×11; (d) row and column offset 50×50; (e) cut 75 lines; (f) rotate 30°

Download full size

Fig. 8. Security experiment. (a) Initial value sensitivity test; (b) chaotic sequence uniqueness test

Download full size

Table 1. Characteristic stability of DCT coefficients

View table

Table 1. Characteristic stability of DCT coefficients

Image	C（1，1）	C（1，2）	C（1，3）	C（1，4）	C（1，5）	C（1，6）	C（1，7）	C（1，8）
A	2.7991	0.2099	-0.8241	-0.0947	0.0854	-0.0317	-0.0683	0.0461
B	3.4839	0.1653	-0.6004	-0.0865	0.0689	-0.0276	-0.0468	0.0430
C	3.1684	0.1891	-0.7450	-0.0833	0.0784	-0.0397	-0.0547	0.0399
D	2.7991	0.2099	-0.8241	-0.0947	0.0854	-0.0317	-0.0683	0.0461
E	2.8049	0.2114	-0.8293	-0.0966	0.0868	-0.0323	-0.0695	0.0467
F	2.7990	0.2099	-0.8241	-0.0947	0.0854	-0.0317	-0.0683	0.0461
G	2.7863	0.1923	-0.8404	-0.1089	0.0678	-0.0560	-0.0908	0.0265

Table 2. Invisibility of watermark
View table
Table 2. Invisibility of watermark
Image name Fundus 1 Fundus 2 Fundus 3 Brain map 1 Brain map 2 Brain map 3
Watermarked image
PSNR /dB 40.7834 40.8793 40.7596 41.0008 41.1943 40.9259

Table 3. Experimental results of non geometric attack on double watermark

View table

Table 3. Experimental results of non geometric attack on double watermark

Attack type	Parameter	BER1	NC1	BER2	NC2
Gaussian noise	0.002	0.0034	0.9966	0.0073	0.9935
Gaussian noise	0.005	0.1130	0.8841	0.1262	0.8864
Salt & Peppernoise	0.002	0.0424	0.9582	0.0424	0.9623
Salt & Peppernoise	0.005	0.0916	0.9094	0.0945	0.9152
Speckle noise	0.005	0.0000	1.0000	0.0000	1.0000
Speckle noise	0.010	0.0000	1.0000	0.0000	1.0000
Median filtering	3×3	0.0000	1.0000	0.0078	0.9930
	7×7	0.0000	0.9996	0.0273	0.9755
	11×11	0.0505	1.0000	0.0818	0.9268
Wiener filtering	3×3	0.0000	1.0000	0.0000	1.0000
	7×7	0.0010	0.9498	0.0125	0.9889
	11×11	0.0706	1.0000	0.0876	0.9213
Gaussian filtering	3×3	0.0000	0.9991	0.0039	0.9965
	7×7	0.0000	0.9301	0.0039	0.9965
	11×11	0.0000	1.0000	0.0039	0.9965
JPEG	20	0.0000	1.0000	0.0117	0.9895
JPEG	30	0.0000	1.0000	0.0000	1.0000

Table 4. Experimental results of geometric attack

View table

Table 4. Experimental results of geometric attack

Attack type	Parameter	BER1	NC1	BER2	NC2
Scaling	0.25	0.0007	1.0000	0.0039	0.9965
	0.5	0.0007	1.0000	0.0000	1.0000
	2	0.0003	1.0000	0.0000	1.0000
	4	0.0160	1.0000	0.0000	1.0000
Crop	Upper left 1/8	0.0003	1.0000	0.0000	1.0000
	Upper left 1/4	0.0007	0.9723	0.0985	0.9119
	Center 1/8	0.0169	0.9981	0.1533	0.8608
	Center 1/4	0.0723	0.9927	0.1992	0.8201
Rotate	15	0.0033	0.9954	0.0085	0.9924
	30	0.0029	0.9983	0.0056	0.9950
	45	0.0091	0.9989	0.0011	0.9990
	90	0.0003	1.0000	0.0000	1.0000

Table 5. Experimental comparison results

View table

Table 5. Experimental comparison results

Attack type	Parameter	Reference［8］	Reference［9］	Reference［10］	NC1	NC2
Gaussian noise	0.002	0.8900	0.8226	0.7634	0.9966	0.9935
Gaussian noise	0.005	0.8546	0.8002	0.7181	0.8841	0.8864
Salt & Pepper noise	0.002	0.9450	0.9602	0.6859	0.9582	0.9623
Salt & Pepper noise	0.005	0.9338	0.9580	0.6973	0.9094	0.9152
Speckle noise	0.002	0.9466	0.9826	0.6800	1.0000	1.0000
Speckle noise	0.005	0.9452	0.9815	0.6663	1.0000	1.0000
Median filtering	3×3	0.5019	0.9346	0.6815	1.0000	0.9930
Median filtering	5×5	0.3993	0.6873	0.8127	0.9996	0.9755
Wiener filtering	3×3	0.8636	0.9662	0.8390	1.0000	1.0000
Wiener filtering	7×7	0.8649	0.7822	0.7692	0.9498	0.9889
Gaussian filtering	3×3	0.8489	0.9522	0.8757	0.9991	0.9965
Gaussian filtering	7×7	0.8005	0.9408	0.8206	0.9301	0.9965
JPEG	20	0.4747	0.7346	0.7658	1.0000	0.9895
JPEG	30	0.6838	0.9315	0.9243	1.0000	1.0000
Scaling	2	0.9350	0.9777	0.8136	1.0000	1.0000
	4	0.9398	0.9791	0.6840	1.0000	1.0000
	0.5	0.8558	0.9544	0.8136	1.0000	1.0000
	0.25	0.4314	0.6169	0.6801	1.0000	0.9965
Crop	Upper left 1/8	0.9466	0.9823	0.6944	1.0000	1.0000
	Upper left 1/4	0.9346	0.9788	0.7067	0.9723	0.9119
	Center 1/8	0.9468	0.9693	0.6770	0.9981	0.8608
	Center 1/4	0.9468	0.9176	0.6669	0.9927	0.8201
Rotate	15	0.8381	0.9339	0.8610	0.9954	0.9924
	30	0.8517	0.9491	0.8541	0.9983	0.9950
	45	0.9246	0.9736	0.8267	0.9989	0.9990
	90	0.9468	0.9823	0.9400	1.0000	1.0000

Table 6. Running time of watermarking algorithm

View table

Table 6. Running time of watermarking algorithm

Algorithm	Technology	Embedding time /s	Extraction time /s	Extraction time /s
Reference［8］	DCT+DFT	1.2634	0.5485	1.8092
Reference［9］	DWT+SVD	1.6563	0.8289	2.4582
Reference［10］	CT+DCT+SVD	0.7390	0.2343	0.9733
Proposed algorithm	DWT+DCT+Schur	0.5570	0.0654	0.6224

Tools

Get Citation

Copy Citation Text

Miaomiao Wang, Deyang Wu, Sen Hu, Jiayan Wang, Yan Wang, Haibo Jin, Changbo Qu. Double Watermarking Algorithm for Tamper Detection of Medical Images[J]. Laser & Optoelectronics Progress, 2023, 60(6): 0617001

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites