An Efficient Resource Utilization Scheme for Video Transmission over Wireless Sensor Networks

IJSRD - International Journal for Scientific Research & Development| Vol. 1, Issue 6, 2013 | ISSN (online): 2321-0613
All rights reserved by www.ijsrd.com 1274
Abstract— In this paper we propose an energy efficient
video transmission strategy for wireless sensor networks,
which combines wavelet-based image decomposition and
cooperative communication. The proposed scheme uses the
selective decode and forward (SDF) cooperation, so that a
relay node collaborates with the source by forwarding only a
lower-resolution version of the original video, obtained via
discrete wavelet transform (DWT). We show that the
proposed SDF-DWT strategy is more energy efficient than
non-cooperative single-hop and multi-hop, also
outperforming the regular SDF scheme. In addition, we
show that our method can achieve the energy efficiency of
incremental DF (IDF), without the need of a feedback
channel.
Keywords: Cooperative communication, video transmission,
wireless sensor networks, DWT.
I. INTRODUCTION
The advancement in wireless communications and
electronics has enabled the development of low power
sensor networks. The sensor networks can be used for
various application areas (e.g., health, military, home). For
different application areas, there are different technical
issues that researchers are currently resolving. Wireless
Sensor Networks (WSNs) are a collection of tiny nodes,
consisting of sensing, data processing and communicating
components, often connected in an ad hoc manner. These
nodes are typically battery-operated and their recharge or
replacement may be undesirable or not possible [1]. Thus,
energy consumption is a key issue in the design of WSNs.
An important WSN class, the Visual Sensor Networks
(VSNs), has a camera as the sensor component and thus deal
with images and videos. A video has an amount of data
many orders greater than typical sensors data, like
temperature, pressure, and humidity. Video data
transmission consumes much more energy than for other
typical WSN applications [2].
The large amount of data traffic for video transmissions over
a WSN can rapidly shorten the lifetime of the network, so a
strategy to mitigate this problem is to promote some sort of
image data compression. Recent works have investigated
transmission in WSNs, and many of them employ wavelet-
based image compression. The [3] authors consider a non-
cooperative multi-hop scenario and propose two schemes
based on discrete wavelet transform (DWT). Semi-reliable
transmission is employed, enabling priority based packet
discarding by intermediate nodes according to their battery
charge. A source node responds to a video query by the
destination node, which specifies the desired video quality.
Videos coming from different paths are combined at cluster
heads, and then multiple copies of the combined videos are
retransmitted. In [5] an energy efficient point-to-Point
transmission scheme is discussed. The DWT is used in order
to obtain different quality layers, and an algorithm
adaptively determines the coding and transmission
parameters.
Considering that network nodes are frequently single
antenna devices, due to cost and size constraints, one
technique to achieve energy savings is cooperative
communications [6], [7]. In this technique a relay node
overhears the communication between the source and
destination nodes and collaborates forwarding the source
information to the destination. One of the most employed
cooperative protocols is known as selective decode-and-
forward (SDF), where the relay only forwards the messages
it can correctly decode. Another cooperative protocol,
incremental decode-and-forward (IDF), is similar to SDF,
but the relay forwards the messages correctly decoded only
when required by the destination node.
Cooperative communications can save energy by reducing
the required transmit power, but the additional RF circuitry
consumption of the relay nodes must be taken into account
[8]. It is important to note that in cooperative schemes the
communication takes place in two time slots, while in case
of non-cooperative single-hop transmission it usually occurs
in a single time slot. So, the transmit power reduction
obtained in cooperative schemes comes at the expense of the
spectral efficiency seen at the receiver. A fair energy
consumption comparison between cooperative and non-
cooperative schemes should consider the same target packet
loss rate and end-to-end throughput. The results in [9] show
that, when these two constraints are considered, the
cooperative schemes can be considerably more energy
efficient than multi-hop and single hop, even for short
transmission distances. The most energy efficient scheme is
IDF [9], but it requires a feedback channel.
In this paper we propose an efficient video transmission
strategy exploiting the multi-resolution characteristic of
DWT. Differently than [3]-[5] , we consider the SDF
cooperative transmission scheme, in which the relay sends a
lower resolution version of the video originally transmitted
from the source to the destination node. Thus, even if the
destination does not receive the video from the source, there
is a probability that it receives it from the relay. As we
assume that transmissions are orthogonal in time and that
the relay transmits less data than the source, the main
contribution of this paper is to show that the overall energy
consumption can be considerably reduced, with a negligible
decrease on the average video quality. Moreover, we show
that the proposed DWT-based SDF scheme can achieve the
energy efficiency of IDF, with the advantage that in our
proposed scheme a return channel is not required. The rest
of this paper is organized as follows. Section II describes the
An Efficient Resource Utilization Scheme for Video Transmission
over Wireless Sensor Networks
K. Swathi1
K. Murali Krishna2
G.S.Naveen Kumar3
Prof. P. Sanjeeva Reddy4
B.Suresh Babu5
1
ECE JNTUH 2,4
Member of IEEE 3
Research Scholar, JNTUK

An Efficient Resource Utilization Scheme for Video Transmission over Wireless Sensor Networks
(IJSRD/Vol. 1/Issue 6/2013/001)
DWT and a video quality assessment metric. Section III
analyzes the energy consumption of some transmission
schemes, while the proposed SDF-DWT is presented in
Section IV. Numerical results are shown in Section V and
finally, Section VI concludes the paper.
II. IMAGEQUALITY METRIC FOR DWT CODED
VIDEOS
The wavelet transform has been frequently used in signal
processing applications. Its inherent capacity for multi
resolution representation motivated a quick adoption and
widespread use of wavelets in image-processing
applications [10]. DWT is a process which can be roughly
described as the decomposition of a digital signal by passing
it through two filters, a low-pass filter (L) and a high-pass
filter (H). This process generates two sub-bands, the low-
pass sub-band representing a down sampled low-resolution
version of the original signal, and the high-pass sub-band
representing the residual information of the original signal,
needed for the perfect reconstruction of the original set from
the low-resolution version. Since image is a bi-dimensional
signal, a 2D-DWT is performed. This is achieved by first
applying the L and H filters to the image row by row, then
re-filtering the output column by column by the same filters.
As a result, the image is divided into four sub-bands, the
𝐿𝐿sub-band, containing the image low-resolution version,
and three sub-bands 𝐻𝐿, 𝐿𝐻and 𝐻𝐻, containing horizontal,
vertical and diagonal residual information, respectively [3].
The decomposition procedure can be repeated recursively to
obtain multiple levels of decomposition. The most
commonly used multidimensional 2D-DWT structure
consists of a recursive decomposition of the 𝐿𝐿sub-band,
illustrated in Figure 1, where 2D-DWT was applied twice.
In this paper we adopt the Le Gall 5-tap/3-tap wavelet filters
[3], which are an energy efficient DWT implementation for
image compression, requiring only fixed point operations.
Fig. 1: 2D-DWT applied (a) once and (b) twice.
For a long time, the Mean-Squared Error (MSE), and the
equivalent Peak-to-Signal Noise Ratio (PSNR), have been
the dominant quantitative performance metrics in the field
of signal processing [11]. Even though MSE and PSNR are
simple to calculate, have clear physical meanings, and are
mathematically convenient in the context of optimization,
they are not very well matched to perceived visual quality
[12]. The main reason is that distinct kind of image
distortions can produce very similar MSE/PSNR values,
although their perceived visual quality be quite different.
Due to this, great efforts have gone into the development of
quality assessment metrics that take advantage of known
characteristics of the human visual system (HVS). Some of
these metrics take into account the structural similarity of
the original and the reconstructed images, rather than an
error measurement. One of such a metric is the Structural
Similarity Index (SSIM), based on the assumption that HVS
is highly adapted to extract structural information from the
viewing field. Thus, a measure of structural information
change can provide a good approximation to perceived
image distortion. Like MSE/PSNR, SSIM is a full-reference
metric, in the sense that it compares a distorted version of an
image with its original one. The similarity between both
videos ranges from 0 (none) to 1 (identical). Assuming that
x and y are local image patches taken from the same
location of two videos that are being compared, the local
SSIM index measures the similarities of three elements of
the image patches: the similarity 𝑙(x, y) of the local patch
luminance (brightness values), the similarity 𝑐(x, y) of the
local patch contrast, and the similarity 𝑠(x, y) of the local
patch structure. These local similarities are expressed using
simple, easily computed statistics, and are combined
together to form local SSIM [12] as:
Where 𝜇 𝑥 and 𝜇 𝑦are, respectively, the local sample means of
x and y, 𝜎 𝑥 and 𝜎 𝑦 are, respectively, the local sample
standard deviations of x and y, and 𝜎 𝑥𝑦is the sample cross
correlation of x and y after removing their means. The
parameters 𝐶1 and 𝐶2 are small positive constants based on
the dynamic range of the pixel values that stabilize each
term, so that near-zero sample means, variances, or
correlations do not lead to numerical instability. According
to, it is useful to apply the SSIM index locally rather than
globally. So, the statistics above are computed locally within
a sliding window that moves pixel-by-pixel across the
image, resulting in a SSIM map. This window has size 11
×11 pixels, which is weighted by a circular-symmetric
Gaussian function w = {wi /1, 2, ..., N}, with standard
deviation of 1.5 samples, normalized to unit sum (∑
). The estimates of local statistics𝜇 𝜎 and 𝜎 are then
computed as:
In practice, a single overall quality measure of the entire
image is calculated by considering the mean SSIM
(MSSIM) index:
Where X and Y are the reference and the distorted images,
respectively, 𝑥 and 𝑦 are the image contents at the j-th local
window, and Miss the number of local windows in the
image.

III. ENERGY CONSUMPTION ANALYSIS
We consider three relevant nodes in a WSN – source (S),
relay (R) and destination (D) – where the source tries to
communicate with the destination, and the relay is at an
intermediate position. In addition, we assume that
transmissions are orthogonal in time and that the nodes are
half-duplex. The methods are characterized in terms of
outage probability, and the Nakagami-mdistribution [13] is
employed to model the wireless propagation environment.
Moreover, we assume that the channel is in long-term quasi-
static fading.
For any given node i, transmitting a frame s, the received
signal at node j is given by:
where P is the transmit power, ℎij is a scalar that
represents the unity variance Nakagami-mquasi-static fading
and n𝑖𝑗represents the AWGN vector, with variance 𝑁0/2 per
dimension, where 𝑁0 is the thermal noise power spectral
density per Hertz. The path loss between 𝑖and 𝑗is given by
[14]:
Where𝐺is the total gain of the transmit and receive
antennas, 𝜆is the wavelength and 𝑑𝑖𝑗corresponds to the
distance between nodes 𝑖and 𝑗. Parameter 𝛼is the path loss
exponent, 𝑀𝑙is the link margin and 𝑁 𝑓is the noise figure at
the receiver.
In the following we analyze the energy consumption in
terms of the total energy consumption per bit, where we also
include the power consumed by the RF circuitry for
transmitting and receiving, which impacts the overall energy
consumption especially at short distances. Two non-
cooperative strategies are considered (single-hop and multi-
hop), and two cooperative strategies (SDF and IDF), while a
novel cooperative scheme, called SDF-DWT, is proposed in
Section IV.
A. Single-hop Transmission (SH)
According to the model introduced in [8], the total
consumed energy per bit in a transmission from 𝑆to 𝐷is:
Where 𝑃 𝑃𝐴, represents the power consumed by the power
amplifier in SH transmission, 𝑃 and 𝑃 , which are
respectively the energy consumed by the transmitting and
the receiving circuitry and corresponds to the bit rate in
bits/s.The energy efficiency analysis is performed under the
constraint of an outage probability. In the transmission of a
frame from node 𝑖to node 𝑗, an outage occurs when the
Signal-to-Noise Ratio (SNR) at node 𝑗falls below a
threshold 𝛽= 2Δ
−1 [14], where Δ = /𝐵is the spectral
efficiency and 𝐵is the system bandwidth. In Nakagami-
𝑚fading, the outage probability of the SH transmission is
given by [15]:
Where Γ(𝑎) and Ψ(𝑎, 𝑏) are the complete and incomplete
gamma functions, respectively, and 𝑁=
𝑁 ⋅𝐵is the noise power. At high SNR (low outage region),
the outage probability can be approximated by [15]:
The minimal transmit power for SH is constrained by a
target end-to-end spectral efficiency Δ and a target outage
probability 𝒪∗at the receiver. Thus, by replacing by 𝒪∗,
such that ≤𝒪∗, the minimal transmit power 𝑃∗
is the
power that minimizes (10) [9].
B. Multi-hop Transmission (MH)
In MH, 𝑆 sends a packet to which then forwards it to 𝐷.
The 𝑆-𝐷link is not utilized. Moreover, we assume that is
able to detect if the packet was received correctly or not, and
forwards it to 𝐷 only if the packet was correctly received.
Otherwise, the packet is considered lost.
As mentioned before, the loss in spectral efficiency inherent
to MH (which takes two time slots) reduces the Maximum
end-to-end throughput to half that of SH [9]. Thus, in order
to obtain the same end-to-end throughput, each transmission
requires twice the spectral efficiency of SH. Consequently,
an outage occurs when the received SNR is below a
threshold 𝛽′= 22Δ
−1. Thus, the outage probability for each 𝑖-
𝑗link is:
The overall outage probability for MH is given by the
combination of the outages in the 𝑆- and -𝐷links:
And, similar to SH, the minimal transmit power for MH,
𝑃∗
can be obtained from (12) by replacing by 𝒪∗.
Then, the total consumed energy per bit in MH is:
The first term in (13) corresponds to the energy consumed
by the first transmission, while second term, which occurs
with probability (1− ), corresponds to the energy
consumed by the retransmission. In addition, note that all
terms are divided by 2 , because with the spectral
efficiency multiplied by two, each single transmission is two
times faster.
C. Decode-and-Forward Cooperative Transmission
As in MH, DF schemes also require two time slots to
perform the communication: a broadcast from 𝑆in the first
time slot followed by a retransmission from in the second
time slot. Considering that selection combining [14] is used
at 𝐷, the end-to-end outage probability for the DF protocol,
which includes the outage probabilities of the three links
involved in the transmission, 𝑆-𝐷,𝑆– and -𝐷, obtained
from (11), is:

Which is the same for both SDF and IDF schemes, and from
which we can obtain the minimal transmit power 𝑃∗
, given
an outage target of 𝒪∗.
D. Selective DF (SDF)
In the SDF protocol, cooperates whenever it successfully
decodes the message from 𝑆. Then, the total consumed
energy per bit is:
Where the first term in (15) represents the energy consumed
in the broadcast phase, and the second term represents the
case where is successful in decoding the message from ,
which is retransmitted to 𝐷. Note that the total energy
consumption of SDF has one additional 𝑃 when compared
to MH. That is because in the first time slot both 𝐷and
have to decode the transmission from 𝑆.
D. Incremental DF (IDF)
The IDF cooperative protocol takes advantage of the
availability of a feedback channel. Through this channel, 𝐷
is able to request a retransmission from if the previous
transmission from 𝑆 was not successful. Thus, only
retransmits if required by 𝐷. Then, total consumed energy
per bit is:
Which only differs from (15) by an additional 𝑆𝐷 in the
second term of the equation, indicating that retransmits
only if the message from the source could be perfectly
recovered, and only if the transmission from 𝑆fails.
IV. PROPOSED SCHEME: SDF-DWT
We consider applications where, although high quality
videos are preferred, videos with lower quality are tolerable.
This way, we proposed an SDF-DWT scheme, based on the
classical SDF protocol; however, instead of retransmitting
the entire video, the relay forwards only a lower-resolution
version of it. As mentioned in Section II, when the 2D-DWT
is applied to an image, four sub-bands are produced, with
𝐿𝐿1 sub-band corresponding to a quarter-size version of the
original image. Applying the 2D-DWT again over 𝐿𝐿1 sub
band produces the 𝐿𝐿2 sub-band which is, by its turn, a
quarter-size version of the 𝐿𝐿1 sub-band image,
corresponding to an image which is 1/16 times smaller than
the original one. This decomposition procedure could be
done many times as desired, but at the expense of a lower
image quality. The resulting total consumed energy per bit
in the proposed SDFDWT is:
Where 𝑁DWT corresponds to the number of times the 2D-
DWT is applied to the original video. The first term in (17)
represents the consumed energy when the entire video sent
by 𝑆is received at and 𝐷, while the second term represents
the case where the lower-resolution video is forwarded by .
Note that, since the time spent during the second time slot is
smaller than in the other schemes by a factor of , a
reduction on the overall energy consumption is expected.
In order to quantify the quality of the retransmitted video,
since always forwards a lower-resolution version of the
original video sent by 𝑆, we define an video overall mean
quality index, . We consider “mean quality” in the
sense that we compute a mean value between the number of
times that 𝐷receives the entire video from 𝑆, and the number
of times that 𝐷receives the lower-resolution Video from ,
weighted by their respective MSSIM values. So, the overall
mean quality index for the 𝑁 -th 2D-DWT level is:
𝑀𝑆𝑆𝐼𝑀 ) +𝑀𝑆𝑆𝐼𝑀 (1-
)(1- (18)
Where 𝑀𝑆𝑆𝐼𝑀 is the MSSIM value for the entire
video (equal to 1), and 𝑀𝑆𝑆𝐼𝑀 is the MSSIM value
when the 𝑁 -th 2D-DWT level is applied to the video.
Therefore, the first term in (18) represents the case when
𝐷has successfully decoded the entire video from 𝑆, and the
second term represents the case when the transmission from
𝑆 to 𝐷 failed, but 𝐷could successfully decodes the video
from .
V. NUMERICAL RESULTS
In this section we numerically compare the energy
efficiency of the presented transmission schemes. We
assume typical parameters used in literature and the simple
WSN scenario mentioned on section III, with three nodes –
source (𝑆), relay ( ) and destination (𝐷). We assume that the
link margin and the noise figure are 𝑀 = 40 dB and
𝑁 = 10 dB, respectively, the total antenna gain is 𝐺= 5 dBi,
the carrier frequency is 𝑓𝑐= 2.5 GHz and 𝑁0 = −174
dBm/Hz. Moreover, we consider a bandwidth of 𝐵= 10 kHz
and the path loss exponent 𝛼= 2.5. Figure 2 shows the total
energy consumed per bit for SH, MH, SDF, IDF, and the
proposed SDF-DWT scheme, for 𝑁 = 1, 𝑁 = 2 and
𝑁 = 3 in the case of NLOS. We assume a maximum
packet loss rate of 𝒪∗= 10−2 and spectral efficiency of Δ = 2
bits/s/Hz. We also assume that is at the midpoint in a
straight line between 𝑆and 𝐷. As we observe, the proposed
SDF-DWT scheme performs much better than single-hop,
multi-hop schemes and, as expected, overcomes SDF in the
whole distance range. The performance almost achieves the
IDF one when 2D-DWT is applied once, and outperforms it
when 2D-DWT is applied two and three times. We show a
similar LOS scenario in Figure 3. Although SH presents a
general improvement compared to the NLOS scenario, SDF-
DWT outperforms all the other schemes when 2D-DWT is
applied one, two and three times.
However, the gains in Figures 2 and 3 come at a potential
loss in video quality. In order to determine the video quality
loss we utilize a 512×512 bitmap gray-scale Yogitha video
as the original test video.
We first apply the 2D-DWT to an image of video generating
the 𝐿𝐿1 sub-band.

Then we perform it recursively two more times, generating
𝐿𝐿2 and 𝐿𝐿3 sub-bands.
Fig. 2: Energy consumption in a NLOS scenario, for 𝒪∗= 10
−2 and Δ = 2 bits/s/Hz.
Fig. 3: Energy consumption in a LOS scenario, for 𝒪∗= 10−
2 and Δ = 2 bits/s/Hz.
Each of these sub-bands is compared to the original image
using a Matlab implementation of MSSIM index algorithm
available online at [16]. In fact, as the images must be equal
in size to be compared by MSSIM, the inverse 2D-DWT is
applied to 𝐿𝐿𝑖sub-bands until the original size is obtained. In
this case, residual information sub-bands 𝐻𝐿𝑖, 𝐿𝐻𝑖and
𝐻𝐻𝑖are all made equal to zero. The obtained MSSIM values
are 0.9301, 0.8324, for 𝐿𝐿1, 𝐿𝐿2 and respectively, which are
very consistent with perceived visual quality of the images,
as can be seen in Figure 4. It is worth to point out that even
though applying 2D-DWT several times produces smaller
images, visual quality also degrades. Besides, the overall
energy consumption does not decrease at the same ratio,
since energy consumption reduction occurs only in the
second transmission time slot. For example, the application
of the third 2D-DWT causes a substantial decrease in the
image quality, from 0.8324 in MSSIM values, as shown in
image (c) of Figure 4.However the energy savings per bit
are very small, as shown in Figures 2 and 3. Hence, a trade-
off between energy savings and video quality must be
considered, depending on the available energy of the nodes
and the video quality application constraints.
In terms of the video overall mean quality index, we show in
Table I the results for considering three maximum
packet loss constraints, 𝒪∗= 10−2, 𝒪∗= 10−3, and 𝒪∗= 10−4.
The more restrictive the packet loss is, the greater the
contribution from 𝑆is in the index. Also, the
contribution from 𝑆is more significant in an NLOS
Environment than when some LOS is present.
Fig. 4: (a) Original image, MSSIM = 1 (b) 𝐿𝐿1, MSSIM =
0.9301 (c) 𝐿𝐿2, MSSIM = 0.8324 .
Table 1: Overall Video Quality Index
VI. CONCLUSION
In this paper we investigate the energy efficiency of a video
transmission in a simple WSN scenario combining wavelet
multilevel decomposition with an SDF cooperative scheme.
The energy consumption for a video transmission is less in
Wireless Sensor Networks by combining the Wavelet
multilevel decomposition with an SDF cooperative
operation. Hence the proposed scheme is more energy
efficient than the Single Hop and the Multi Hop technique.
REFERENCES
[1] L. F. Akyildiz,W. Su, Y. Sankarasubramaniam, and
E. Cayirci, “A survey on sensor networks,” Computer
Networks, vol. 38, no. 4, pp. 393– 422, Jan. 2002.
[2] S. Soro and W. R. Heinzelman, “A survey of visual
sensor networks,” Advances in Multimedia, vol.
2009, p. 21, 2009.
[3] V. Lecuire, C. Duran-Faundez, and N.
Krommenacker, “Energy-efficient video transmission
in sensor networks,” International Journal of Sensor
Networks, vol. 4, pp. 37–47, Jul. 2008.
[4] H. Wu and A. A. Abouzeid, “Error resilient video
transport in wireless sensor networks,” Computer
Networks, vol. 50, pp. 2873–2887, Oct.
2006.
[5] W. Yu, Z. Sahinoglu, and A. Vetro, “Energy efficient
Avi video transmission over wireless sensor
networks,” IEEE Global Telecommunications
Conference (GLOBECOM), vol. 5, pp. 2738–2743,
2004.
[6] A. Sendonaris, E. Erkip, and B. Aazhang, “User
cooperation diversity - part I: System description,”

IEEE Trans. on Communications, vol. 51, pp. 1927–
1938, 2003.
[7] J. N. Laneman, D. N. C. Tse, and G. W. Wornell,
“Cooperative diversity in wireless networks: efficient
protocols and outage behavior,” IEEE Trans. Inform.
Theory, vol. 50, pp. 3062–3080, 2004.
[8] S. Cui, A. J. Goldsmith, and A. Bahai, “Energy-
constrained modulation optimization,” IEEE Trans.
on Wireless Communications, vol. 4, pp. 2349–2360,
2005.
[9] G. Brante, M. Kakitani, and R. Souza, “Energy
efficiency analysis of some cooperative and non-
cooperative transmission schemes in wireless sensor
networks,” IEEE Trans. Commun., vol. 59, no. 10,
pp. 2671– 2677, Oct. 2011.
[10] J. Fowler and B. PesquetPopescu, “An overview on
wavelets in source coding, communications, and
networks,” EURASIP Journal on Image and Video
Processing, vol. 2007, 27 pages, 2007.
[11] Z. Wang and A. C. Bovik, “Mean squared error: Love
it or leave it? A new look at signal fidelity measures,”
IEEE Signal Processing Magazine, vol. 26, pp. 98–
117, 2009.
[12] Z. Wang, A. C. Bovik, H. R. Sheikh, and E. P.
Simoncelli, “Video quality assessment: From error
visibility to structural similarity,” IEEE Trans. on
Video Processing, vol. 13, no. 4, pp. 600–612, 2004.
[13] M. K. Simon and M.-S. Alouini, Digital
Communication over Fading Channels. Wiley
Interscience, 2004.
[14] A. Goldsmith, Wireless Communications, 1st ed.
Cambridge University Press, 2005.
[15] Z. Wang and G. Giannakis, “A simple and general
parameterization quantifying performance in fading
channels,” IEEE Trans. on Communications, vol. 51,
no. 8, pp. 1389 – 1398, aug. 2003.
[16] Z. Wang, “The ssim index for image quality
assessment,” http://www.cns.nyu.edu/ lcv/ssim/.

An Efficient Resource Utilization Scheme for Video Transmission over Wireless Sensor Networks

More Related Content

What's hot (20)

Viewers also liked (8)

Similar to An Efficient Resource Utilization Scheme for Video Transmission over Wireless Sensor Networks (20)

More from ijsrd.com (20)

Recently uploaded (20)

An Efficient Resource Utilization Scheme for Video Transmission over Wireless Sensor Networks