Digital Video Watermarking Based on 3D-Discrete
Wavelet Transform Domain
Sadik. A.M .Al-Taweel
#1
, Putra, Sumari
*2
, Hailiza Kamarulhaili
School of Computer Sciences, Mathematical School, Universiti Sains Malaysia, Minden
11800, Penang, Malaysia
#3
1 Sadik@cs.usm.my
3 hailiza@cs.usm.my
2 putras@cs.usm.my
Abstract—One of the significant problems in video
watermarking is the Geometric attacks. The DWT (discrete wavelet
transform) domain is used for proposed a novel algorithm to place
invisible watermark in a video frame based on a three-level DWT
using Haar filter. The proposed algorithm is robust against JPEG
compression, geometric attacks such as Downscaling, Cropping,
and Rotation. It is also robust against Image processing attacks
such as low pass filtering (LPF), Median filtering, and Weiner
filtering. Furthermore, the algorithm is robust against Noise
attacks such as Gaussian noise, Salt and Pepper attacks. The
embedded data rate is high and robust. The experimental results
show that the embedded watermark is robust and invisible. The
watermark was successfully extracted from the video after various
attacks.
I. INTRODUCTION
It is important for Digital watermarking to have digital data
and multimedia, such as video, music, text, and image
copyright protection because of network and multimedia
techniques that easily copy. Digital watermarking is a general
solution that can be used to identify illegal copying and
ownership, authentication, or other applications by inserting
information into the digital data in visible, or an invisible way
[1].
In order for a watermark to be useful, it must be
perceptually invisible and have robustness against a variety of
possible attacks and image processing by those who seek to
corsair the material [2], [,3]. There has been much emphasis
on the robustness of watermarking against signal processing
operations. However, it has become clear that a very small of
the geometric distortions can prevent the detection of a
watermark in many watermarking techniques. There are three
advantages of using watermarking [4]. First, the watermark is
embedded into the media content in an imperceptible way, so
that the information embedded is hidden and the effect on the
appearance and function of the work is invisible.
Second, the watermark is embedded in the content of the
data and it is supposed to survive even after encryption and
decryption, convert from digital to analog, compression,
change of file format, resizing, or cropping.
Third, any changes made can be identified by checking the
watermark because it is embedded in such away as to locate
the modified parts and restore the original.
One of the challenges in the video and image watermarking
is the geometric manipulations of the watermarked data. It is
known that a small amount of rotation or scaling can
dramatically disable the receiver from detecting the
watermark. Geometric transformations are yet to be solved
because of their high computation complexity,
implementation difficulties, and poor performance [5]. So,
geometric manipulation will destroy the synchronization of
the watermarking embedding and detection process. The
detection of the watermark requires a synchronization step to
locate the embedded watermark in the content. Some of the
video watermarking techniques are based on the spread
spectrum approach which targets the geometric distortions
and operation in the 3D wavelet domain and BCH codes, and
3D interliving by incorporating an effective temporal
synchronization technique [6], [ 7].
Reviews on many other techniques that transform the video
watermarking into discrete wavelet transform domain to
handle the geometric distortions, attacks, and applications can
be found in [8], [9] . In [10], the proposed technique works in
the space domain and it is robust against geometric
transformations like image cropping and resizing. Serdean et
al [11] Presented robust geometric attack techniques such as
shift, rotation, scaling and cropping in the spatial domain by
employing image registration techniques. The weakness of the
existing algorithms includes the low bit rate of the watermark.
Other published papers in this domain are Chan et al [12],
the researchers presented a novel DWT-based video
watermarking scheme with scrambled watermark and error
correcting code. The scheme is robust against attacks such as
frame dropping, frame averaging, and statistical analysis.
Campisi et al [13] proposed perceptual mask, applied in the
3D DWT domain robust against MPEG2 and MPEG-4
compression, collusion and transcoding attacks.
In [14] propose a robust mpeg video watermarking in
wavelet domain which they embedded in two bands (LL and
HH) and chosen attacks JPEG compression, resizing, adding
Gaussian noise, low pass filtering.
In [15] present a novel adaptive watermarking scheme
based on error correction code and Human Visual System
(HVS) in 3D-DWT domain.
The proposed method is to resist signal processing attacks,
Gaussian noise, and frame dropping.
In [16] propose a method based on 3D wavelet transforms.
In this method the original video frames are divided into 3D-
block according to HVS properties. The proposed method is
2009 IEEE International Conference on Signal and Image Processing Applications
978-1-4244-5561-4/09/$26.00 ©2009
352