Advancements in science and technology have introduced the need to protect data, authenticate data, integrate data, assert ownership, content labelling and security. Digital Watermarking schemes protect all forms of digital data. Digital Image Watermarking can be applied to gray scale, halftone, color, medical and 3D images. The process of watermarking can be broadly classified into three phases namely embedding, attacking, and decoding for typical scenarios. Some of the watermarking schemes adopted in the past include vector quantization, spread spectrum, SVD, DCT, DFT, etc. It was observed that the spread spectrum was more robust and it had also been applied for patenting. In spite of this, the method could not withstand high amplitude noise. Hence, later DCT, DFT and Wavelets were used. These schemes were not robust to collusion attacks. In this review, we have identified the embedding and detection schemes of the existing watermarks over the past decade and analyzed the robustness of each of these methods. The different parameters considered to analyze the performance of the existing watermarking schemes are also discussed. Research under watermarking is a great field of interest involving multimedia security, forensics, data authentication and digital rights protection. This paper will be useful for researchers to implement a robust watermarking scheme.

1. Digital Image Watermarking Techniques: A Review
Pushpa Mala .S. pushpasiddaraju@gmail.com
Research Scholar, Jain University, Bengaluru, India
& Sambhram Institute of Technology, Bengaluru, India
D. Jayadevappa devappa.22@gmail.com
Dept. of Electronics & Instrumentation Engineering,
JSS Academy of Technical Education, Bengaluru, India
K.Ezhilarasan murali981983@gmail.com
Research Scholar, Jain University, Bengaluru, India
& Sambhram Institute of Technology, Bengaluru, India
Abstract
Keywords:
1. INTRODUCTION

2. 2. FEATURES OF DIGITAL WATERMARKING SCHEMES
Robustness
Imperceptibility
Security
Verifiability
Computational Cost
Watermark Detection
Capacity
Tradeoff parameters

3. FIGURE 1:
3. DIGITAL IMAGE WATERMARKING APPROACHES
3.1. Visible and Invisible Watermarking
et.al
3. 2. Private and Public Watermarking
et al.
, was .
et al.

4. also
3.3. Robust and Fragile Watermarking
N N k
A
k l, P Q
et al.
ing scheme
the scheme was
3.4. Blind and Non-blind Watermarking
.
Nonblind watermarking schemes access
the host image during the detection phase and are private in nature. et al.
schemes
et al.
et al.
3.5 Spatial and Frequency Domain Watermarking
.
4. FREQUENCY DOMAIN WATERMARKING APPROACHES
prior
ample
s

5. was
was
was was
of could be
n N N N N
n
et al.
et al.
8 8
* , , , , , ,
, ,
, ,
C
u v b u v b u v b
u v b
u v b
X t w
X X

 

, , , ,if
otherwise
c
u v b u v bX t
, ,u v bX *
, ,u v bX
, ,u v bw , ,
c
u v bt
, , , , , , ,*
, , ,
, , ,
F
u v l f l f u v l f
u v l f
u v l f
X t w
X X

 

, , , ,if
otherwise
F
u v l f l fX t
, , ,u v l fX  ,u v l
f ,
*
, , ,u v l fX , , ,u v l fw
,
F
l ft l
f
et al. scheme
T

6. scheme
et al.
the watermark
scheme
scheme
8 8
lc- lc
et al.
lc lc
et al.

7. et al.
N N
 
1 2
1 1 1 2 2 2
1 2
1 1
2 / 2 /
1 2 1 2
0 0
, ( , )
N N
j n k N j n k N
n n
I k k i n n e  
 
 
 
  
1 2 1 2( , ) ( , )M k k I k k 1 2( , )P k k
1 2( , )W k k
A N N U V
S W S
WU WV WS ‘ AW
U WS T
WV
 T
A U S V
  T
W W WS αW U S V
A  T
W WU S V

8. et al.
et al.
et al.
 
2
2 /2
1/22 2
( ) (2 ,
where
x
x x e
x x y


 
 
2
2* /2
( )
k k
k k e


 k
     , .i j i jP k M x M x 
   2 (2 . *i i
iM x x A 

( )iM x , ( )i jP k
i j

9. 






et al.
M≥2
l l
al

10. et al.
et al.
et al.
et al.
.
5. EVALUATING WATERMARKS
5.1.Performance Evaluation Metrics
PSNR
N N
MSE


11. BER
C
H W
Transform
Adopted
Concept & Details Results & Summary
SS
DCT
DFT
DWT
Multiwavelets
Contourlets
Ridgelets
Curvelets
Shearlets
TABLE 1:
SSIM
p,q
  
  
1 2
2 2 2 2
1 2
2 2 pqp q
p qp q
c c
c c
  
   
 
  
 p
 q
 p q respectively

2
p p

2
q

12. 
2
pq p q
 1c 2c
MSE
MSE =  
2
1 1
1
( , ) ( , )
p q
i j
N i j W i j
pq  
 
 
  

p q
( , )N i j
( , )W i j
5.2.Watermarking Attacks
et al.
1) Compression Attacks
2) Interference Attacks
3) Signal Processing Attacks
4) Geometric Attacks
5) Cryptographic Attacks-
6) Collusion Attacks-
7) Active Attacks –

13. 8) Passive Attacks
9) Histogram Equalization attacks
Attack No Attack Median
Filtering
Salt and
Pepper
Noise
Gaussian
Filter
PSNR
TABLE 2:
6. CONCLUSION
7. REFERENCES