This full-text paper was peer-reviewed and accepted to be presented at the IEEE WiSPNET 2016 conference.

A Hybrid Approach for Video Steganography using

Edge Detection and Identical Match Techniques
Ramandeep Kaur,Pooja,Varsha
CT Group of Institutions
Shahpur Jalandhar, India
raman177rk@gmail.com , poojachoudhary80@gmail.com , barkhabright@gmail.com

AbstractTo maintain the secrecy of information, various

information hiding techniques are used. Steganography is
one of them, which means hide information inside another
digital media like text, image, audio, video etc. without
being detected by human visual system (HVS). Recently
Video steganography has become a strong tool to hide
large amount of data rather than image steganography.
This paper is proposing an idea about a hybrid approach
for video steganography to achieve high capacity data &
high quality of stego video on the basis of quality metrics
like PSNR, MSE and BER. The proposed methodology is a
combo pack of various techniques such as RSA encryption,
Edge detection, Identical Match and 4LSB substitution, in
which a video file will be used to hide text message inside
in all layers of RGB color frames of video. The
experimental results are analyzed on MATLAB software
on cover video of rhinos and text secret message is
embedding inside the video. The resulting values show that
our proposed algorithm has high imperceptibility and
security and resists the visual attacks.
Keywords Video Steganography; 4LSB; Identical Match;
RSA Encryption; Canny Edge Detector; Secret Key.

I. INTRODUCTION
Nowadays security of private information is a major issue
over the internet. Because in todays digitized world, the
whole communication is done through internet and
transferring private data from one end to another using various
applications such as e-mails, chats, etc. but there is main issue
that is how to protect our confidential information from
hackers or cyber criminals over internet. To solve such
problems and to maintain the security of data, we should
follow a algorithm which should not only encrypt the data into
another form but also hides its presence and video
steganography helps to provide such a secure environment
over internet. To protect the private information from being
misused by the attackers and to overcome the alteration of
information, a novel data hiding approach is used.
Steganography is one of the best information hiding technique,
which hides the presence of secret message behind a
multimedia file without changing the perceptual quality of

c
978-1-4673-9338-6/16/$31.00 2016
IEEE

media file and provide secure communication between two

parties. On the basis of Digital media, steganography is
categorized as; text, image, audio, video and protocol based
steganography. Here we are dealing with video steganography.
Video steganography is a process of hiding the secret
information behind video bit streams. The main goal of video
steganography is to hide presence of secret message from
human visual system. Various companies and organizations
are following this concept to secure their confidential
information and databases from intruders. Video files can hide
large amount of hidden data behind their bit streams than
images. So, that why they are more preferable than image
steganography.
II. PROBLEM STATEMENT
We have studied that there is lots of limitations in
previous algorithms which are not good enough for video
steganography process. In previous research works [7], the
symmetric key based encryption algorithms (like XOR
Transformation, permutation operation, DES, 3DES and AES)
are used which can be easily decrypted by the attackers to
understand the code of original hidden message instead of
Asymmetric algorithms. As, sequential encoding is a good
encryption method for securing the embedding process of
video steganography but still it has some limitations as by
using sequential encoding, intruders can identify the presence
of hidden message by sequentially analyzing the video frames.
So this technique is not much secure for steganography
process. To preserve the security of secret information, RSA
encryption algorithm and key based random frame selection
algorithms are followed in our proposed work.
For embedding process, 1LSB technique is used to
embed data in video frames. But simple LSB technique is not
resist to attacks and produces the poor value of peak signal to
noise ratio (PSNR) and high value of mean square errors
(MSE) which are both the image quality parameters, so to
avoid this problem we will use an improved version of this
technique. [4] Hiding in LSB bits change the resolution of
pixels which may be easily visible to human eye during
detection mechanisms and also they are very easy t attack by
attackers. . But proposed technique can embed large amount of
data than previous one and shows the better result for video
quality parameter. To maintain the quality of video, identical

896

This full-text paper was peer-reviewed and accepted to be presented at the IEEE WiSPNET 2016 conference.

match based substitution algorithms are applied in our

proposed work. Our algorithm provides better solution for
social problems such as manipulation of data and frame
dropping. As we are hiding same secret message in all selected
random frames. For example, if any unauthorized person such
as hacker or cyber criminals tries to drop the some
steganographed frames then we can still recovered the secret
information from other left stegano frames. So it helps to resist
attacks on video clip.
For edge detection process, various edge detectors are
used such as sobel, Robert, prewit, laplacian etc. but some
of these edge filters has some limitations due to which they
are not good enough to find better and true edges from
selected media. As Sobel operator is not enough accurate to
find out edge pixels from image area. Robert filter has
limitation to give some false edges and some dashed lines.
And laplacian operator increases the complexity and
processing time as it uses 2nd derivative for computation. In
our work, to increase the security of video, secret key is
used which verifies the owner of video. In extraction time,
if a user enters 3 times wrong secret key then his access will
be denied due to unauthorized access and whole video file
will get damaged and corrupted by leaving no secret
message behind it.
III. PROPOSED METHODOLOGY
In proposed work a novel and hybrid approach of edge
detection and identical match will be used for secret message
inside a different video. The proposed methodology is a
combination of various techniques and will help to achieve
better results than the previous work had been done as shown
in Fig.4. It helps to increase capacity and security of data.
A. Phases of Proposed Work
The main phases for proposed methodology are following:
a) Pre-Processing: In our research work, we are selecting
cover video (.avi file) as a cover video and fragment this video
into number of frames (nof). These frames act as a carrier
medium to hide secret message.
b) Random Frame Selection: In this phase, we are
selecting the 10 random frames from extracted frames of video
using secret key. Because, If we are extracting frames
sequentially, then attacker can easily identify the presence of
secret messages inside the video frames but random frames
make it difficult to analyze.
c) Edge Detection: In our work, we are embedding secret
message behind edge pixels as well as non edge pixels. The
edge detection process highlights the edge regions from given
image. The edges are used to hide secret information because
edges are very sharp in nature and their pixels intensity values
are changed very frequently and have discontinuity effect.
They can be brighter or dimmer & can tolerate variations in
pixel values changes due to high contrast or color values and if
we change modify these color values then human eye cant see
changed pixels in between the edges and there is no
degradation in quality of media file. So, that why we have
preferred both edge pixels & smooth areas to hide secret data.

To detect the edge pixels of 10 random selected frames, we are

using canny edge detector. Canny edge detector is an optimal
operator because of good detection, good localization and
single respond to an edge. [10] It helps to detect wide range of
edges & in these edge pixels we can embed large amount of
pixel values of secret data and remove the noise present in an
image than other detectors. It uses multistage algorithm to
detect wide range of edges. So we can hide high capacity data
in edge pixels. For detecting edge pixels from all pixels of
selected frame, we are using an algorithm in which if selected
pixel value is 1 then it is edge pixel, otherwise it is non- edge
pixel.
d) RSA Encryption: To make our secret message more
secure, we are using RSA asymmetric key based encryption
algorithm. It converts the secret message in an unreadable
format and increases its security from hackers. In our work,
we have used RSA encryption algorithm instead of symmetric
key based encryption. The reason behind this is as the
symmetric key based algorithms such as XOR encryption,
AES, DES etc. are not secure enough because secret key is
always shared with third party members and their security
depends on the trustworthy nature of member, and can be
easily breakable by attackers. But there is no need of third
party member to share secret key between sender and receiver
because both sender and receiver are using their own public
and private keys to encrypt and decrypt the secret message and
also used and large factorial number for encryption, which is
hard to decrypt by attackers. So RSA encryption provides high
security to our algorithm for hiding the encrypted secret
message behind video frames.
e) Embedding Process: In our work, we are Hiding the
secret message (.text) using 4LSB substitution method in
between the edge pixels and using 2 bit identical match (ID)
technique is used to hide secret bits in between the smooth
areas (Non edge pixels).
f) Extraction Process: To achieve the final output of our
algorithm, we have decoded the hidden information from
stego video and display the secret message as an output.
g) Comparison: In our work, we are comparing
traditional LSB and proposed methodology on the basis of
payload capacity, PSNR, MSE, and BER and achieving high
imperceptibility performance and high hiding capacity and
security for video steganography process.
IV. ALGORITHMS
A. Data Hiding Algorithm (Transmitting Side)
At sender side, sender selects video file from browsing
window and hide secret text behind random frames using
following algorithm:
x Select cover video (.avi) & fragment it into different
frames.
x Enter secret key and select random frames as per
secret key based functions and calculate slab size
using Eq. 1.

897

This full-text paper was peer-reviewed and accepted to be presented at the IEEE WiSPNET 2016 conference.

Slab-size =
(1)

x
x
x
x

x
x

V. EXPERIMENTAL RESULTS & PERFORMANCE

CALCULATIONS

F1 = Secret-Key(d1) + Secret-Key(d2);
F2 = Secret-Key(d3) +Secret-Key(d4);
..
F10 = Secret-Key(d11) + Secret-Key(d12);
Where, d1,d2,d3,d4,d5,d6,d7,d8,d9,d10,d11,d12 are the
secret key digits ACSII values and F1to F10 are the
selected random and dynamic frames for hiding the
secret message bits. [12]
Enter text secret message (.txt) and encrypt it using
RSA algorithm.
Detect edges of selected frames using canny edge
operator as shown in Fig.5.
Hide secret key, selected frames addresses and
message length in RED (R), GREEN (G), BLUE (B)
components of first fixed frame respectively.
Now, Hide 4 bits of secret message in places of edge
pixels and non edge pixels using 4LSB substitution in
blue layer and 7pair based identical match technique
respectively.
In 7 pair identical match technique; firstly we will
check identical pair of secret message bits from 7 pair
identical values of bits in video frame pixels. If there
is existence of identical pair bits in non edge pixels
then we will note bit position in raw table. Otherwise
hide bits using 2LSB technique. Here we are hiding 2
bits in green and red layer each.
Compute quality metrics for all selected frames such
as MSE, PSNR, BER and histogram analysis.
Recombine, all the frames to generate a stego video
file and transfer it over network using any
communication media such as e-mail etc.

B. Extraction Algorithm (Receiver Side)

At receiver side, secret message is extracted from stego
video file. The following steps are followed for extraction
process:
x Select stego video and fragment it into number of
frames to extract the secret message from frames.
x Extract selected frame addresses, secret key and
message length from the first fixed frame of stego
video.
x Take secret key from user for authentication process
and compare it with originally entered or extracted
key. If success, than continue for extraction process.
Otherwise, video get damaged due to unauthorized
access after 4 attempts.
x Decrypt secret message by using RSA decryption
algorithm.
x Display secret message as final output.
x Exit.

We are using MATLAB 7.10.0 for implementing the proposed

methodology and for achieving the experimental values for
algorithm. Here, video file rhinos.avi is used as a cover file
which have 15 fps frame rate and 114 total number of frames
(nof). 10 random frames are selected for hiding the text (.txt)
secret message of length 2048 bytes (2 KB) and each frame is
reshaped into size 512*512. The secret key is used
1234567890 which is used for selecting the 10 random
frames from video clip as well as used to extract the secret
message from stego video file. We are hiding 4 bits in each
pixel of frame to increase the capacity of hidden data in a
video frame, edge of frame and corresponding stego frame is
shown in Fig.1.
A. Hiding Capacity
The maximum hiding capacity of our proposed algorithm is
128 KB. Such as total number of bits can be hidden = n*m*
no. of bits= 512*512*4 hidden bits per pixel=1048576 bits; in
bytes = 1048576/8= 131072 bytes; & in KB=
131072/1024=128KB (Maximum Hiding Capacity). We are
hiding 4 bits per pixel of frame. So we are achieving high
embedding capacity from our experimental results. The
resulting values of all hiding capacity such as 128 bytes to 64
KB are tabulated in Table 1 which shows the comparison
between previous work [2] and proposed work on the basis of
capacity.
B. Imperceptibility Performance Calculations
Imperceptibility is a measure of quality metrics. The quality
metrics such as PSNR (Peak to signal noise ratio), MSE
(Mean Square Error), and BER (Bit Error Rate) are calculated
for all selected frames to analyze the performance of
imperceptibility. Figure 2 & 3 are showing the PSNR, BER
values for all selected frames and respectively. The
experimental results show that our algorithm has high PSNR
and low MSE & BER values which shows high
imperceptibility performance.
C. Security
The proposed algorithm protects the secret message from
visual and statistical attacks due to identical bit matching
substitution mechanism. It has high security than the previous
algorithms due to random frame selection on the basis of
secret key. Also RSA encryption algorithm provides one layer
more protection to our algorithm and only those user can
decrypt the secret message that have correct secret key,
otherwise stego video file will get damaged because of
unauthorized access for video steganography process.
VI .CONCLUSION & FUTURE WORK
Video steganography is a very active research field with
lots of applications and preferred as an information hiding
tool. The proposed methodology is a hybrid approach, which
is the combination of edge detection, identical match, 4LSB
substitution and RSA encryption algorithms. We are achieving
high embedding capacity i.e. 128 KB data and getting high

898

This full-text paper was peer-reviewed and accepted to be presented at the IEEE WiSPNET 2016 conference.

quality and security because of high PSNR and low BER

values. Using these methodologies, we have obtained better
results for hiding large amount of data. In future work, we can
use different formats of video such as .mov, .mp4, .Flv to hide
secret message behind video and also different format of text
can be hide behind video such as text in Hindi, Punjabi, &
Arabic language etc.

Pre-Processing (Fragmentation of
Video)

Random and Dynamic Frame

Selection using Secret Key

Edge Detection Using Canny Edge

Detector

Encryption of Secret Message Using

RSA Algorithm

Embedding Process
Fig. 1. Selected Video Frame, Edge of Frame and Stego Frame
Edge
Pixels

Non-Edge
Pixels

Using 4 LSB
Substitution

Using Identical
Match Technique

Extraction Process

Fig .4. Flow Chart of Proposed Methodology

Edg
e-R
R
Fig. 2. PSNR values for all Stego Frames (For 2 KB Data)

Select
ed
Fram
es

Edge
Detectio
n Using
Canny
Edge
Detector

Edg
e- G

OR
Operat
or
(|)

Edg
e-B

EdgedFrames

FIG.5. EDGE DETECTION ALGORITHM

Fig. 3. BER values for all Stego Frames (For 2 KB data)

899

This full-text paper was peer-reviewed and accepted to be presented at the IEEE WiSPNET 2016 conference.

International Journal of Hybrid Information Technology Vol.6, No.5 ,

2013, pp.59-66.
[10] Soumyajit Sarkar, Arijit Basu, Comparison of various Edge Detection
Techniques for maximum data hiding using LSB Algorithm,
International Journal of Computer Science and Information
Technologies, Vol. 5, Issue 3, 2014.
[11] Youssef Bassil, Image Steganography based on a Parameterized
Canny Edge Detection Algorithm, International Journal of Computer
Applications, Vol. 60, No.4, December 2012.

TABLE I. COMPARISON OF MSE, PSNR, AND BER ON THE BASIS OF DATA

HIDING CAPACITY
Hiding
Capacit
y in
bytes
128
bytes
2048
bytes(2
KB)
4096
bytes(4
KB)
16384
bytes(16
KB)
32768
bytes
(32KB)
65536
bytes
(64KB)

Previous Work

Proposed Work

MSE

PSNR(db)

BER

MSE

PSNR(
db)

BER

9.3361
e-004
0.0124

78.44

0.0127

96.98

0.0104

67.19

0.0149

3.2976
e-005
0.0051

75.11

0.0134

0.0249

64.17

0.0156

0.0108

71.65

0.0141

0.0996

58.15

0.0172

0.0547

63.45

0.0159

0.1986

55.15

0.0181

0.1337

57.92

0.0173

ERROR
Capacity
Exceeded

ERROR

ERRO
R

0.2987

53.55

0.0173

[12] Amrinder Singh, Sukhjit Singh, A Robust Video Watermark

Embedding and Extraction Technique Based on Random Frame
Selection, IJRIT International Journal of Research in Information
Technology, Vol. 2, Issue 2, February 2014, pp: 28-37.

ACKNOWLEDGMENT
The authors would like to thank to the previous work
for proposing different mechanisms for video steganography
which contributes a lot for work made in this paper. All work
done in this paper will surely help to the new researchers for
doing future work in video steganography field.
REFERENCES
[1]

[2]

[3]

[4]

[5]

[6]

[7]

[8]

[9]

Ramandeep Kaur, Pooja and Varsha, The Non Tangible Masking of

Confidential Information using Video Steganography, International
Journal of Computer Applications (IJCA), Vol.119, No.17, June 2015.
Ramandeep Kaur, Pooja, XOR Encryption Based Video
Steganography, International Journal of Science and Research (IJSR),
Vol.4, Issue 11, November 2015.
Hemant Gupta, Setu Chaturvedi, Video Steganography Through LSB
Based Hybrid Approach, International Journal of Computer Science
and Network Security, vol.14, No.3, pp 99-106, March 2014.
Sunil. K. Moon, Rajeshree. D. Raut, Analysis of Secured Video
Steganography Using Computer Forensics Technique for Enhance Data
Security, IEEE Second International Conference on image
information processing (ICIIP- 2013), pp 660-665.
Geetha C.R., H. D. Giriprakash, Image Steganography by Variable
Embedding and Multiple Edge Detection using Canny Operator,
International Journal of Computer Applications (0975 888) ,Vol. 48,
No.16, June 2012.
Mamta Juneja and Parvinder Singh Sandhu, Improved LSB based
Setganography Techniques for Color Images in Spatial Domain,
International Jounal of Network Security, Vol.16, No. 6, pp- 452-462,
Nov.2014.
Pooja Yadav, Nishchol Mishra, Sanjeev Sharma, A Secure Video
Steganography with Encryption Based on LSB Technique, IEEE
International Conference on Computational Intelligence and
Computing Research, 2013.
Atallah M. Al-Shatnawi, A New Method in Image Steganography with
Improved Image Quality, Applied Mathematical Sciences, Vol. 6,
2012, no. 79, pp .3907 3915.
Sarabjeet Kaur and Sonika Jindal, Image Steganography using Hybrid
Edge Detection and First Component Alteration Technique,

900