secure and distributed video surveillance via portable devices
TRANSCRIPT
ORIGINAL RESEARCH
Secure and distributed video surveillance via portable devices
Pietro Albano • Andrea Bruno • Bruno Carpentieri •
Aniello Castiglione • Arcangelo Castiglione • Francesco Palmieri •
Raffaele Pizzolante • Kangbin Yim • Ilsun You
Received: 3 December 2012 / Accepted: 21 March 2013
� Springer-Verlag Berlin Heidelberg 2013
Abstract In this work we present a system for distributed
video surveillance based on the Client-Server model. The
system we present can be accessed via portable devices. In
many real-world scenarios is useful, or sometimes neces-
sary, to have portable devices that can receive real-time
data from a selected camera, to prevent or to manage
anomalous activities. The system provides reliable, high
speed, secure and real-time communication among all its
components, which are the Repository, the Node and the
Portable Device. Both Repository and Node can act as a
server. The Repository can provide services to both Nodes
and Portable Devices, while the Nodes provide services
only to the Portable Device. The portable device can only
act as a client, using the services offered by the other two
parts. In our system, a portable device is assumed to know
only the location of the Repository which permits to get the
list of nodes connected with one or more camera(s). When
a portable device gets the list, it can choose which node
intends to connect to, to get the images of its connected
camera(s). The security of the interaction among Node-
Repository and Node-Portable Device is guaranteed by
using the SSL/TLS protocol. The interaction among nodes
and portable devices is secured by using an invisible digital
watermarking algorithm on each image, before that image
is sent from a node to a portable device. The latter extracts
the watermark from the image and verifies the identity of
the node.
Keywords Remote video surveillance � Remote personal
security � Mobile video surveillance � Homeland security �Real time communication � Security and privacy
enforcement
1 Introduction
Video surveillance is today an important and increasingly
used tool for the monitoring of areas and environments
P. Albano � A. Bruno � B. Carpentieri � A. Castiglione �A. Castiglione � R. Pizzolante (&)
Dipartimento di Informatica, Universita degli Studi di Salerno,
84084 Fisciano, SA, Italy
e-mail: [email protected]
P. Albano
e-mail: [email protected]
A. Bruno
e-mail: [email protected]
B. Carpentieri
e-mail: [email protected]
A. Castiglione
e-mail: [email protected]; [email protected]
A. Castiglione
e-mail: [email protected]
F. Palmieri
Dipartimento di Ingegneria dell’Informazione,
Seconda Universita di Napoli, 81031 Aversa, CE, Italy
e-mail: [email protected]
K. Yim
Department of Information Security Engineering,
Soonchunhyang University, Asan, Korea
e-mail: [email protected]
I. You
School of Information Science, Korean Bible University,
Seoul, Korea
e-mail: [email protected]
123
J Ambient Intell Human Comput
DOI 10.1007/s12652-013-0181-z
(Muller-Schneiders et al. 2005; Norris et al. 2002). Video
surveillance is commonly used in many real-world sce-
narios and can be applied both in civil (Srinivasan et al.
2004; Norris et al. 2002; Collins et al. 2000) and military
(Huang et al. 2009) fields. For example, important appli-
cations of video surveillance can be found in the identifi-
cation of individuals and objects (Ko 2008) as well as in
the prevention and detection of other abnormal activities
(Nasution and Emmanuel 2007; Duque et al. 2007). Video
surveillance is also helpful in other fields such as agricul-
ture (for the prevention of fires Foresti and Snidaro 2002;
Toreyin et al. 2005, 2006; Chen et al. 2003, 2004), traffic
monitoring (Tseng et al. 2002), etc.. Video surveillance
brings many advantages with respect to classical surveil-
lance, for instance it reduces the number of people who
work for surveillance and therefore it reduces the resources
and the costs that are necessary to security. Through a
video surveillance system it is possible to detect and also to
record activities that could be useful for offline analysis
and, in some cases, the recorded information that can be
used as evidence of crimes (Nieto 1997).
It is well-known that portable devices have changed our
lives giving us a handy tool for several and different uses.
Moreover, their adoption introduced several security issues
Castiglione et al. (2009) that have been dealt with by
introducing applications making use of cryptographic
primitives (De Santis et al. 2010; Castiglione et al. 2011;
Castiglione et al. 2012).
On the contrary, from a positive point of view, with the
diffusion of portable devices having an enhanced power of
computation and advanced features, it is now possible to
perform surveillance and monitoring activities directly
through these devices. This feature can be very useful for
example when traveling or, in general, if someone needs to
control an environment when moving frequently from one
place to another.
In this work we propose a secure distributed system for
video surveillance based on the Client-Server model which
provides the opportunity of remote connections using por-
table devices, for the sake of real-time monitoring. A pre-
liminary version of this work appears in (Albano et al.
2012). The system architecture is composed by three main
components: a central server (Repository), one or more
collector Node(s) and one or more Portable Device(s). The
Repository has a list of collector nodes which are connected
with one or more camera(s). The Portable Devices who want
to access the system have to know only the location (host-
name or IP address) of the Repository. When the connection
is established, the Repository sends to the Portable Device a
list of nodes, so that this device can choose which Node it
intends to connect to, and, after this choice is made, the
device connects to the Node and gets the image frames or
any other multimedia contents obtainable by the camera(s).
The remainder of this work is organized as follows: in
Sect. 2 we focus on the proposed distributed architecture.
In Sect. 3 we analyze the security issues of the proposed
system and we illustrate our decisions regarding the design
and the development of the prototype we made. In Sect. 4
we describe the interaction among the system and the end-
user, and, finally, in Sect. 5, we present our conclusion and
highlight our future work directions.
2 The distributed client-server architecture
The Repository mainly deals with the localization of the
Nodes, it interacts with them and with the Portable Devices
which want to access the system. For this reason, the
Repository holds a list of all Nodes that already joined the
system.
A collector Node is connected with one (ore more) cam-
era(s) and interacts with the Portable Devices interested in
monitoring the areas controlled by that Node. The Portable
Device has to know only the location (IP address or host-
name) of the Repository, which gives them a list of all the
available collector Nodes as well as other descriptive infor-
mation about them. When a Portable Device has joined the
system, it receives, at fixed intervals, the images obtained by
the selected camera(s) connected to the selected Node.
Figure 1 shows an overview of the proposed system, in
particular it illustrates the system architecture.
2.1 Interaction Node–Repository
The interaction between a Node and the Repository takes
place through a TCP connection over a SSL/TLS tunnel.
The use of these protocols is fundamental for the security
Fig. 1 An overview of the proposed system architecture
P. Albano et al.
123
of our system, since, without a secure (encrypted and
authenticated) connection, a ‘‘fake’’ Node could easily join
the system and send manipulated/tampered images to the
incoming Portable Devices. The messages exchanged
between a Node and the Repository are of three types:
Login, Register, and Disconnect. These commands are
encoded through XML sequences.
Figure 2 shows a complete exchange of messages dur-
ing the interaction among the above parties, including each
of the three messages.
The Node uses the Login command for authenticating
itself to the Repository. If the authentication succeeds, the
Node gets as a response the ‘‘200’’ return code, otherwise,
the Repository denies the access for that Node. Figure 3
shows the XML structure representing an example of Login
command sent from a Node to the Repository.
The Node uses the Register command for sending its
descriptive information to the Repository. If the authenti-
cation succeeds, the Node gets an ID from the Repository
which uses this ID to uniquely identify the Node. The ID is
also used by the Node during its disconnection process
from the Repository. Figure 4 shows an example of Reg-
ister command sent from a Node to the Repository.
When a Node wants to disconnect from the Repository,
it sends a Disconnect command. Figure 5 shows an
example of Disconnect command sent from a Node to the
Repository.
2.2 Interaction Repository–Portable Device
As for the interaction between a Node and the Repository,
also in those between the Repository and the Portable
Devices, it is used a TCP connection over a SSL/TLS
tunnel.
The interaction among these two parts occurs after the
invocation of one of the following three commands by the
Portable Device: ListServer, Update and Desc. By using
the ListServer command the Portable Device requests the
Repository to download the list of all Nodes currently
registered. The reply of the Repository can be either a
‘‘530’’ return code in case of error, or a string that has the
following structure:
where:
n is the number of entries;
ts is a timestamp used to indicate the time at which the
request has been sent, it is used with the Update command,
to check if the list is updated;
IDi is a unique identifier associated to node
i (1 B i B n);
Namei is the alias or the name of the node i (1 B i B n);
IPAddressi is the IP address of the node i (1 B i B n).
The Portable Device uses the Update command to check
if its local list of Nodes is updated. The Update command
takes as parameter the timestamp, that is used by the
Repository to verify that the list of Nodes maintained by the
Portable Device is updated. The Repository can reply with
a ‘‘400’’ return code if there are no updates, or otherwise it
replies with a string structured as follows:
where:
n is the number of the entries (n);
Fig. 2 A graphical representation of a complete interaction between
a Node and the Repository. From left to right are shown the
commands (in orange) sent from a Node to the Repository, from rightto left are shown the responses (in light blue) from the Repository to
the Node (color figure online)
Fig. 3 The XML structure of the Login command
Fig. 4 The XML structure of the Register command
Fig. 5 The XML structure of the Disconnect command
Secure and distributed video surveillance via portable devices
123
ts is a timestamp which is used for indicating when the
request has been sent;
1IDi is the node with the identifier IDi that has been
registered after the last request of update, or
2IDi is the node with the identifier IDi that has been
disconnected after the last request of update;
Namei is the alias or the name of node i (1 B i B n);
IPAddressi is the IP address of the node i (1 B i B n).
2.3 Overhead reduction for repository–portable devices
communication
For the reply messages from the Repository to the Portable
Devices the use of textual strings instead of XML
sequences has been adopted. This choice permits to reduce
the communication overhead among these two communi-
cating parts.
Consider, as an example, the Fig. 6, and suppose that
this is the XML sequence of the minimum length that could
be used by the Repository to reply the Portable Device.
In Table 1 is shown the number of additional bytes
introduced by using the XML sequence of Fig. 6 compared
to the usage of a textual string that represents the same
information. Clearly, these additional bytes can be seen as
an overhead. In this example, we assume that each char-
acter costs one byte, and that also in the XML sequence
there are no blank spaces or other tabulation characters.
The first column shows the number (n) of Nodes connected
to the Repository, the second and third column show the
number of overhead bytes needed respectively by using the
XML sequence and the textual strings. The fourth column
indicates the difference in percentage between the two
types (XML and textual strings) of overhead.
Figure 7 shows a graphical representation of the over-
head trend with a large number of Nodes, the red dotted
line represents the overhead trend of the minimal length
XML sequences, while the blue one represents the trend of
the textual strings.
2.4 Interaction Node–Portable Device
The communication among these two parts takes place via
TCP socket, using an FTP-like protocol for the exchange of
both commands and data. The main difference between the
standard FTP and the one introduced by the authors, is that
in the former the client can open two communication
channels (one for the messages and one for the data), while
in the latter a Portable Device, connected via a GPRS/
UMTS connection, cannot open more than one communi-
cation channel because of the restrictions and policies
imposed by mobile phone operators.
In our ad-hoc protocol, the Server (Node) opens two
sessions, each of them having its own communication
Table 1 Trend of the overhead between minimal length XML
sequences and textual strings
Numberof nodes
XML(overhead)
Textual(overhead)
Differencepercentage
1 42 8 78.57
2 63 13 79.37
3 84 17 79.76
4 105 21 80.00
5 126 25 80.16
6 147 29 80.27
7 168 33 80.36
8 189 37 80.42
9 210 41 80.48
10 231 45 80.52
Fig. 6 An example of a possible minimal length XML sequence
Fig. 7 Graphical comparison of the overhead (in bytes) on Y-axis, in
relation to the number of Nodes (on X-axis) between the messages
sent by using XML minimal length sequences approach (red dottedline) and textual messages approach (blue line) (color figure online)
P. Albano et al.
123
channel, while the Client (Portable Device) opens only a
single channel. The Node communicates to the Portable
Device the port of the other opened channel, and in this
way we can resemble to the standard FTP bypassing the
restrictions imposed by mobile phone operators. The three
commands used during the interaction between these two
parts are Login, List, and MGet.
The Portable Device uses the Login command for
authenticating itself to the Node. The Login operation can
be successful or not. On success, the Portable Device gets
from the Node a message containing the ‘‘200’’ return code
and the interaction among these two parts continues nor-
mally. On failure, the Portable Device gets an alert mes-
sage on its display.
The List command is used by the Portable Device to
request a snapshot (which is stored as a JPEG compressed
image) of all the environments monitored by the Node to
which it is connected. When a Node receives this command
it takes a snapshot from each of its connected cam-
era(s) and send it to the Portable Device.
Finally, the Portable Device uses the MGet command to
request the monitoring of a particular area that is identified
by an univocal code. When the Node receives this com-
mand it creates a data channel by which it can send mul-
timedia frames at fixed intervals.
3 The system security architecture
The design and development of architectures and distrib-
uted protocols that can guarantee security in video sur-
veillance is a challenging issue as shown in Zhang et al.
(2005) and Dufaux et al. (2006). With the generic term
‘‘security’’ we intend controlled and authenticated access to
the system, privacy protection and authentication among
the parts constituting the system (see Liu et al. (2005)).
In order to achieve a secure system with respect to the
security definitions given above, several solutions in the
literature have been proposed, and all of them are based on
the use of cryptosystems. The development of secure
architectures, providing controlled access, privacy protec-
tion, content confidentiality and authenticity, is one of the
most challenging issues in the video surveillance area, and
several solutions, based on the use of cryptography have
been proposed (e.g., Castiglione et al. 2011, 2012; Fleck
and Straßer 2010).
Furthermore, the existence of a surveillance system
strongly depends on legal boundaries (Hunker and Probst
2011) that states ‘‘what’’ is allowed to be monitored, what
is not, and also ‘‘who’’ is authorized to perform monitoring.
In these cases, the data produced by surveillance activities
must be properly secured against unauthorized accesses or
misuses of the collected images.
To define the security of our system, we have to con-
sider all the possible interactions among the parties, and
therefore the various security issues that may arise. For this
reason, in the current section we focus on the Node–
Repository interaction and subsequently on that on between
Node and Portable Device. In the first case of interaction
(and also in that between Repository and Portable Device)
security is guaranteed by using the SSL/TLS protocols, so
that the communication across a network is secured against
eavesdropping and tampering. The same protocol can be
also used for strong cryptographic mutual authentication.
However, there may still be Portable Devices that for
efficiency or hardware limitations, can support only par-
tially (or not support at all) the cryptographic primitives and
protocols which constitute the basis of SSL/TLS. For this
reason, we decided to use digital watermarking techniques
to guarantee a sufficient level of security during the inter-
action among a Node and a Portable Device, even when the
latter has a (very) limited computational power. In partic-
ular, to ensure that an image is not tampered, before sending
the image, each Node includes in such image a digital
invisible watermark. When the Portable Device receives the
image, it extracts the watermark in order to verify whether
the image has been tampered or not.
3.1 Digital watermarking to improve the security
of the proposed system
Digital watermarking is one of the most commonly used
techniques to insert and hide data into digital contents.
When a signal is protected by using a robust digital
watermark, then the associated hidden information will be
also included in all its copies. Watermarking is also fre-
quently used to prevent unauthorized copy of digital media.
There are different embedding methods to include hid-
den data into digital contents, such as, for example Spread-
Spectrum (Liang and Ding 2008; Wang et al. 2000; Bender
et al. 1996), or Amplitude Modulation (Kutter et al. 1998).
In the spread-spectrum method, the signal affected by
the digital watermark is obtained by using an additive
modification. Also in the amplitude modulation the marked
signal is obtained by using the same type of modification
used in the spread-spectrum embedding method, but in this
case the watermark is only embedded in the spatial domain.
Before sending the images, each Node embeds a digital
invisible watermark. When the Portable Device receives
the image, it extracts the watermark in order to verify that
each image has not been tampered.
The watermarking algorithm used in our system has
been proposed by Pizzolante and Carpentieri (2012), and it
is based on a modified version of the one proposed by
Langelaar et al. (1996). It takes four inputs: the source
image, the watermark string, a seed and a threshold T. The
Secure and distributed video surveillance via portable devices
123
watermark string is converted into a bit matrix, in which
each character is converted in a 5 9 8 sub-matrix of bits
(an example is reported in Fig. 8). The resulting string of
bits is obtained by reading the bit matrix line-by-line from
the left-top corner. The seed represents an ID (such as a
numeric PIN) which is used to embed the watermark into
the image, and then to extract it from the watermarked
image. The last parameter is a threshold T, which is a real
number indicating the robustness of the watermark that will
be embedded into the image.
We use the following algorithm to embed a digital
invisible watermark into an image:
1. Convert the image from the RGB domain to the
YUV domain.
2. Convert the watermark string into a bit matrix.
Convert each into a 5 9 8 matrix of bits (see the
example in Fig. 8). The resulting matrix will be
embedded into the original image line-by-line from
the left-top corner.
3. Select, in a pseudo-random way, a block B of 8 9 8
pixels from the image, to embed one bit of the
watermark string.
4. Generate a fixed binary pseudo-random pattern of the
same size of B.
5. Calculate the I0, I1 and D quantities from B. I0 and I1
are obtained by calculating the averages of the
luminance values in B, respectively where the
random pattern is 0 and where the random pattern
is 1. D is the difference between I1 and I0 (D = I1 -
I0).
6. Calculate B0, a reduced quality block obtained from B
by applying on it the quantization and the 8 9 8 DCT
(Discrete Cosine Transform).7. Calculate the I0
0, I1
0and D0 quantities from B0. I0
0and I1
0
are obtained by calculating the averages of the
luminance values in B0, respectively where the
random pattern is 0 and where the random pattern
is 1. D0 is the difference among I0
0and I0
^0
(D^0 = I0
0- I0
^0).
8. If the bit to embed has value 1 then go to step 10.
9. In order to embed the bit value 0, if D and D0 are
greater than the threshold T, subtract the binary
pseudo-random pattern from the block B. The steps
Fig. 10 The repository GUI
Fig. 11 The node GUI
Fig. 8 Example of conversion from the character ‘e’ to the matrix of
bits composed by 5 9 8 (40*bits). The white cells are represented by
the value 0 and the black cells are represented by the value 1
Fig. 9 a The original ‘‘Lena’’ image; b The ‘‘Lena’’ image affected
by a digital invisible watermark
P. Albano et al.
123
6–8, and 10 are repeated iteratively until both
differences are less or equal than -T. Go to step 12.
10. In order to embed the bit value 1, if D and D0 are less
or equal than the threshold T, add the binary pseudo-
random pattern to the block B. The steps 6–8, and 11
are repeated iteratively until both differences are
greater than T.
11. Apply the steps from 4 to 11 to all the pseudo-
randomly selected blocks until all the bits of the
watermark string are embedded.
12. Convert the image in YUV domain back to the RGB
domain.
In Fig. 9a and b are shown respectively the original
‘‘Lena’’ image and the ‘‘Lena’’ image which embeds a digital
invisible watermark of 200 bits (the string ‘‘SeCAM’’)
obtained applying the previously described algorithm of
watermarking. It is easy to see that there are no perceptible
(by the human eye) differences between the two images.
Analogously, we use the following algorithm to extract
the digital invisible watermark:
1. Convert the image from the RGB domain to the YUV
domain.
2. Select, in a pseudo-random way, a block B of 8 9 8
pixels from the image to read one bit of the watermark
string.
3. Generate a fixed binary pseudo-random pattern of the
same size of B.
4. Calculate I0, I1 and D from B. I0 and I1 are then
obtained by calculating the averages of the luminance
values in B, respectively where the random sequence is
0 and where the random sequence is 1. D is the
difference between I1 and I0 (D = I1 - I0).
5. If D [ 0, then the embedded bit has value 1 else it has
value 0.
4 The end-user interface
From the end-user perspective, the system we propose is
composed by three basic components, each of one is rep-
resented in a graphical user interface (GUI) and allows the
interaction between an end-user and the system. These
components are:
– The Repository GUI (described in Sect. 4.1)
– The Node GUI (described in Sect. 4.2)
– The Portable Device GUI (described in Sect. 4.3)
These components are designed and developed in order
to make the interaction between the end-user and the sys-
tem as user-friendly as possible. Moreover, several checks
are performed about the correctness of the entered inputs.
With these checks, the possibility that the end-user enters
incorrect or malformed data is strongly reduced.
4.1 The Repository GUI
The GUI of the Repository is very intuitive and simple. It is
composed by two panels: the first one permits to start the
discovery service provided by the Repository to the Por-
table Devices and to the other Nodes that want to join the
system, while the second one allows to configure the
information about the Repository when it acts as a Node
Fig. 12 a Shows an example of the first step of the Portable Device GUI; b shows an example of the Portable Device GUI with the list of the
nodes; c shows an example of the image obtained by the selected camera of the Node
Secure and distributed video surveillance via portable devices
123
(i.e., the username, the password and so on). An example of
the Repository GUI is shown in Fig. 10.
4.2 The GUI of a Node
The GUI of a Node allows an end-user to enter five
information: the username and the password (obtained
during the registration process), that will be used by each
Node for authenticating itself to the Repository, the Name
or the alias, that will be used by the Repository to identify
the Node, the hostname (or the IP address) of the Reposi-
tory, and, finally, a description of the Node.
Figure 11 shows an example of the Node GUI.
4.3 The Portable Device GUI
This GUI allows the Portable Device to join the system and
to use the services that it provides. The interaction with the
entire system takes place in two phases. In the first one
(Fig. 12a) the end-user have to input the information
regarding the connection with the Repository. In more
details, the information required are the following: user-
name, password and host. These information are used by
the Portable Device to connect and authenticate itself to
the Repository. In the second one (Fig. 12b) the Portable
Device shows to the end-user the list of Nodes obtained
from the Repository. For each node the portable device
GUI provides the opportunity to invoke three commands:
Details, Update and Connect. The Details command allows
the Portable Device to obtain a detailed description of the
Node. The Update command allows the Portable Device to
request the update of its local list of Nodes. The Connect
command allows the Portable Device to connect with a
specific Node and to obtain the frame sequence captured by
the camera(s) to which that Node is connected. Finally, the
Exit command allows the Portable Device to disconnect
from the system and stop all the operations.
5 Conclusions and future research directions
Nowadays, video surveillance has became an important
and increasingly used tool for monitoring areas and envi-
ronments. Video surveillance is used in many real-world
scenarios and can be applied both to civil and military
fields. With the diffusion of portable devices with enhanced
power of computation and advanced features, it is now
possible to perform surveillance and monitoring activities
directly by using these devices. This features can be very
useful for example when traveling or, in general, in situa-
tions in which is needed to control an environment when
moving from one place to another.
In this work, we propose a secure distributed system for
video surveillance based on the client-server model, which
provides the opportunity of remote connections using
portable devices, for the sake of real-time monitoring.
Future research will consider the migration of the proposed
system architecture from the client-server model to the
peer-to-peer (P2P) model, so in this way we intend to
support the scalability of the system when the number of
nodes grows up. In addition, the above migration will allow
to remove all the ‘‘bottlenecks’’ derived by having a single
server that provides all the services. Consider for example
if, for some reasons, the Repository is not available. In this
case the Nodes that have already joined the system cannot
be discovered by the Portable Device, and also, the
incoming Nodes that would to join the system cannot
perform this operation because the system can be accessed
only through the Repository.
Moreover, using video and audio streaming could allow
the portable device to invoke advanced commands such as:
Play, Pause, Stop, Frame Capture and any other standard
function commonly provided by a ‘‘multimedial’’ operating
systems.
Another future research direction is to include in our
system a Motion Detection Engine, in this way the end-user
could be notified, using several communication methods, if
an anomaly has been detected, even when one is not con-
nected to the system.
Finally, it could be meaningful to consider the possi-
bility to send/transfer from a Node to a Portable Device a
compressed short video with audio.
References
Albano P, Bruno A, Carpentieri B, Castiglione A, Castiglione A,
Palmieri F, Pizzolante R, You I (2012) A secure distributed
video surveillance system based on portable devices. In: CD-
ARES, LNCS 7465, pp 403–415
Bender W, Gruhl D, Morimoto N, Lu A (1996) Techniques for data
hiding. IBM Syst. J. 35(3–4):313 –336 (ISSN: 0018-8670)
Castiglione A, De Prisco R, De Santis A (2009) Do you trust your
phone. In: Noia T, Buccafurri F (eds) E-commerce and web
technologies. Lecture notes in computer science, vol 5692.
Springer, Berlin, Heidelberg, pp 50–61 (ISBN: 978-3-642-
03963-8). doi:10.1007/978-3-642-03964-5_6
Castiglione A, Cattaneo G, De Maio G, Petagna F (2011) Secure end-
to-end communication over 3G telecommunication networks. In:
Proceedings of 5th international conference on innovative
mobile and internet services in ubiquitous computing (IMIS)
pp 520–526. doi:10.1109/IMIS.2011.65
Castiglione A, Cattaneo G, Cembalo M, De Santis A, Faruolo P,
Petagna F, Petrillo UF (2012) Engineering a secure mobile
messaging framework. Comput Secur 31(6):771–781 (ISSN:
0167-4048). doi:10.1016/j.cose.2012.06.004
Chen TH, Kao CL, Chang SM (2003) An intelligent real-time fire-
detection method based on video processing. In: Proceedings of
P. Albano et al.
123
IEEE 37th annual 2003 international Carnahan conference on
security technology, pp 104–111
Chen TH, Wu PH, Chiou YC (2004) An early fire-detection method
based on image processing. In: Proceedings of IEEE ICIP’04,
international conference on image processing, 2004, vol 3,
pp 1707–1710
Collins RT, Lipton A, Kanade T, Fujiyoshi H, Duggins D, Tsin Y,
Tolliver D, Enomoto N, Hasegawa O, Burt P, et al (2000) A
system for video surveillance and monitoring, vol 102. The
Robotics Institute, Carnegie Mellon University, Pittsburgh
De Santis A, Castiglione A, Cattaneo G, Cembalo M, Petagna F,
Petrillo UF (2010) An extensible framework for efficient secure
SMS. In: Proceedings of CISIS, 4th international conference on
complex, intelligent and software intensive iystems, Krakow,
Poland, pp 843–850. doi:10.1109/CISIS.2010.81
Dufaux F, Ouaret M, Abdeljaoued Y, Navarro A, Vergnenegre F,
Ebrahimi T et al (2006) Privacy enabling technology for video
surveillance. Proc SPIE 6250:205–216
Duque D, Santos H, Cortez P (2007) Prediction of abnormal
behaviors for intelligent video surveillance systems. In: Pro-
ceedings of IEEE symposium on computational intelligence and
data mining, CIDM 2007, pp 362–367
Fleck S, Straßer W (2010) Towards secure and privacy sensitive
surveillance. In: Proceedings of the 4th ACM/IEEE international
conference on distributed smart cameras, ICDSC ’10, ACM
2010, New York, NY, USA, pp 126–132 (ISBN: 978-1-4503-
0317-0). doi:10.1145/1865987.1866008.
Foresti GL, Snidaro L (2002) A distributed sensor network for video
surveillance of outdoor environments. In: Proceedings of IEEE
2002 international conference on image processing 2002, vol 1,
pp 1–525
Huang Z, Ren S, Chen Y, Jiang W (2009) Research on the military
video surveillance system with conditional access information.
Inf Technol Manage 10:028
Hunker J., Probst CW (2011) Insiders and insider threats—an
overview of definitions and mitigation techniques. J Wirel
Mob Netw Ubiquitous Comput Dependable Appl (JoWUA)
2(1):4–27
Ko T (2008) A survey on behavior analysis in video surveillance for
homeland security applications. In: Proceedings of IEEE 37th
IEEE workshop on applied imagery pattern recognition,
AIPR’08, pp 1–8
Kutter M, Jordan FD, Bossen F (1998) Digital watermarking of color
images using amplitude modulation. J Electr Imag 7:326–332.
doi:10.1117/1.482648
Langelaar GC, van der Lubbe JCA, Biemond J (1996) Copy
protection for multimedia data based on labeling techniques.
In: Proceedings of 17th symposium on information theory in the
Benelux
Liang Q, Ding Z (2008) Spread spectrum watermark for color image
based on wavelet tree structure. In: Proceedings of international
conference on computer science and software engineering, vol 3,
pp 692–695. doi:10.1109/CSSE.2008.958
Liu Z, Peng D, Zheng Y, Liu J (2005) Communication protection in
IP-based video surveillance systems. In: Proceedings of 7th
IEEE international symposium on multimedia, December 2005,
Irvine, Calif, USA, pp 69–78
Muller-Schneiders S, Jager T, Loos HS, Niem W (2005) Performance
evaluation of a real time video surveillance system. In:
Proceedings of 2nd Joint IEEE International Workshop on
visual surveillance and performance evaluation of tracking and
surveillance, pp 137–143
Nasution AH, Emmanuel S (2007) Intelligent video surveillance for
monitoring elderly in home environments. In: Proceedings of
IEEE 9th workshop on multimedia signal processing, MMSP
2007, pp 203–206
Nieto M (1997) Public video surveillance: is it an effective crime
prevention tool? California Research Bureau, California State
Library Sacramento, CA
Norris C, McCahill M, Wood D (2002) The growth of CCTV: a
global perspective on the international diffusion of video
surveillance in publicly accessible space. Surveill Soc 2(2/3):
110–135
Pizzolante R, Carpentieri B (2012) Copyright protection for images
on mobile devices. In: Proceedings of 6th international confer-
ence on innovative mobile and internet services in ubiquitous
computing (IMIS), pp 585–590. doi:10.1109/IMIS.2012.73
Srinivasan S, Latchman H, Shea J, Wong T, McNair J (2004)
Airborne traffic surveillance systems: video surveillance of
highway traffic. In: Proceedings of the ACM 2nd international
workshop on video surveillance and sensor networks, VSSN ’04,
ACM, New York, NY, USA, pp 131–135. (ISBN 1-58113-934-
9) doi:10.1145/1026799.1026821
Toreyin BU, Dedeoglu Y, Cetin AE (2005) Flame detection in video
using hidden markov models. In: Proceedings of IEEE interna-
tional conference on image processing, ICIP 2005, vol 2,
pp 1230–1233
Toreyin BUT, Dedeoglu Y, Gudukbay UCetin AE (2006) Computer
vision based method for real-time fire and flame detection.
Pattern Recogn Lett 27(1):49–58
Tseng BL, Lin CY, Smith JR (2002) Real-time video surveillance for
traffic monitoring using virtual line analysis. In: Proceedings
of IEEE international conference on multimedia and expo,
ICME’02, vol 2, pp 541–544
Wang YP, Chen MJ, Cheng PY (2000) Robust image watermark with
wavelet transform and spread spectrum techniques. In: Proceed-
ings of conference record of the 34th Asilomar conference on
signals, systems and computers, vol 2, 29 Nov 2000, pp 1846–
1850. doi:10.1109/ACSSC.2000.911307
Zhang W, Cheung SC, Chen M (2005) Hiding privacy information in
video surveillance system. In: Proceedings of the 12th IEEE
international conference on image processing, pp 868–871
Secure and distributed video surveillance via portable devices
123