Threshold based filtering technique for efficient moving object detection and tracking in video surveillance

IJRET: International Journal of Research in Engineering and Technology eISSN: 2319-1163 | pISSN: 2321-7308
_______________________________________________________________________________________
Volume: 05 Issue: 02 | Feb-2016, Available @ http://www.ijret.org 303
THRESHOLD BASED FILTERING TECHNIQUE FOR EFFICIENT
MOVING OBJECT DETECTION AND TRACKING IN VIDEO
SURVEILLANCE
P.Vijayakumar1
, A.V.Senthilkumar2
1
Head, Department of Computer Applications, Sri Jayendra Saraswathy Maha Vidyalaya College of Arts
and Science, Coimbatore, Tamilnadu, India.
p_viji_kumar@yahoo.com
2
Director of MCA, Hindusthan College of Arts and Science, Coimbatore, Tamilnadu, India.
avsenthilkumar@yahoo.com
Abstract
Detection and tracking of moving objects are an important research area in a video surveillance application. Object tracking is
used in several applications such as video compression, surveillance, robot technology and so on. Recently many researches has
been developed for video object detection, however the object detection accuracy and background object detection in the video
frames are still poses demanding issues. In this paper, a novel framework called Threshold Filtered Video Object Detection and
Tracking (TFVODT) is designed for effective detection and tracking of moving objects. TFVODT framework initially takes video
file as input, and then video frames are segmented using Median Filter-based Enhanced Laplacian Thresholding for improving
the video quality by reducing mean square error. Next, Color Histogram-based Particle Filter is applied to the segmented objects
in TFVODT framework for video object tracking. The Color Histogram-based Particle Filter measures the likelihood function,
particle posterior and particle prior function based on the Bayes Sequential Estimation model for improving the object tracking
accuracy. Finally, the objects detection is performed with help of Improvisation of Enhanced Laplacian Threshold (IELT) to
enhance video object detection accuracy and to recognize background moving object detection. The proposed TFVODT
framework using video images obtained from Internet Archive 501(c) (3) for conducting experiment and comparison is made with
the existing object detection techniques. Experimental evaluation of TFVODT framework is done with the performance metrics
such as object segmentation accuracy, Peak Signal to Noise Ratio, object tracking accuracy, Mean Square Error and object
detection accuracy of moving video object frames. Experimental analysis shows that the TFVODT framework is able to improve
the video object detection accuracy by 18% and reduces the Peak Signal to Noise Ratio by 23 % when compared to the state-of-
the-art works.
Keywords: Object segmentation, Object tracking, Object Detection, Enhanced Laplacian Thresholding, Median
Filter, Color Histogram-based Particle Filter
-------------------------------------------------------------------***-----------------------------------------------------------------------
1. INTRODUCTION
Video surveillance systems rely on the ability to detect
moving objects in the video sequence which is a relevant
information mining step in a wide range of computer vision
applications. Many research works has been conducted for
video object detection and tracking. Self Crossing Detection
for Parametric Active Contours (SCD-PAC) [1] was
introduced for tracking objects in real world video
sequences with the help of Sobolev active contours.
However, the object detection accuracy remained
unaddressed. The Automatic Estimation of Multiple Motion
(AEMM) was performed in [2] using motion
correspondence algorithm that automatically detects the
multiple motion of object.
Visual recognition system using local invariant detectors
was developed in [3] for improving the rate of features
being detected with the aid of template-based descriptor. But
it is not suitable for complex articulated objects. Automatic
foreground object detection was presented in [4] using
probabilistic consensus foreground object template for
enhancing the foreground object detection results. However,
foreground object cannot be perfectly detected. Multimodal
analysis [5] for moving objects using Canonical Correlation
Analysis (CCA) was performed to improve the accuracy of
moving objects being detected. However, sounding object
segmentation is not effective in some complex scenarios. An
analysis of moving object detection using different image
registration techniques was presented in [6].
Automatic Estimation of Multiple Motion [7] fields were
performed using motion correspondence algorithm with
aiming at improving the automatic identification of
trajectories. A parallel histogram using particle filters [8]
was designed for improving the robustness of object
tracking. But, this algorithm is not robust to large variations
in the appearance of the target. With the aim of removing
the noise present in color video, fuzzy rules were applied [9]

_______________________________________________________________________________________
that resulted in the minimization of Mean Absolute Error
(MAE) and Peak Signal-to Noise Ratio. In [10], Adaptive
Appearance Model was introduced for video tracking using
application dependent thresholds. However, object tracking
accuracy was not at required level. The object tracking
accuracy is improved in TFVODT framework by using
Color Histogram-based Particle Filter.
The algorithm designed in [11] used fuzzy aggregation for
object localization and robustness with intensity and texture
features. Moving object segmentation method in [12] used
mean shift, kalman and particle filter to improve the
accuracy and robust for contour tracking. A tracking
framework introduced in [13] for tracking the multiple
objects based on Continuous Energy Minimization. In [14],
a new approach was presented for segmenting objects in
videos taken in complex scenes with multiple and different
targets.
A robust segmentation method was introduced in [15] for
efficient tracking of segments among various videos using
least square tracking framework. Existing work on moving
object representation using segmentation [16] takes into
account two dominant color histograms to maintain compact
and efficient object representation model. A robust
algorithm was designed in [17] for automatic, noise
detection and removal from moving objects in video
sequences. Dynamic threshold method was introduced in
[18] for moving object detection and Tracking.
Detecting and tracking robust algorithm [19] was developed
for moving object of intelligent video surveillance system
that optimizes the accuracy while maintaining the
computational complexity low. A study of localizing target
was conducted in [20] to improve detection accuracy. In
[21], ELT Method was introduced for multiple moving
object segmentation in video surveillance. In [22], proposed
a novel Framework for Specific Object Tracking in the
Multiple Moving Objects Environment. In [23], MELT
method was introduced for moving object segmentation in
video sequences. In [24], IELT technique initiates the
process of video object segmentation, object tracking and
finally object detection. Though accuracy was improved, but
the mean square error incurred during object tracking
remained unsolved. The mean square error rate is reduced in
TFVODT framework by applying Median Filter-based
Enhanced Laplacian Thresholding.
Based on the aforementioned techniques and methods
presented, in this work we propose a novel framework
called Threshold Filtered Video Object Detection and
Tracking (TFVODT) for improving the moving object
detection and tracking accuracy in video surveillance.
TFVODT framework initially segment the video objects and
reduce the noise frame by applying Median Filter-based
Enhanced Laplacian Thresholding technique. Next, the
segmented video objects are used for efficient tracking of
objects using the Color Histogram-based Particle Filter.
Finally, we perform an IELT for efficient Back Ground
Object Detection.
The paper is organized as follows. In Section 2, we describe
the design of TFVODT framework using Enhanced
Laplacian Thresholding and filtering techniques. Section 3
introduces the different experimental settings studied in this
work. We evaluate our algorithm for object detection and
compare with the state of the work, discuss in detail in
Section 4. Finally, the paper concludes in section 5.
2. DESIGN OF THRESHOLD FILTERED VIDEO
OBJECT DETECTION AND TRACKING FRAME
WORK USING ENHANCED LAPLACIAN THRES
HOLDING AND FILTERING TECHNIQUES
The structure of Threshold Filtered Video Object Detection
and Tracking technique is described in this section. The
TFVODT framework is implemented under the Enhanced
Laplacian Thresholding and filtering techniques. The
TFVODT framework used two filtering techniques i.e.
Median Filter-based Enhanced Laplacian Thresholding and
Color Histogram-based Particle Filter for improving the
video quality and reducing the noise frame. Improvisation of
Enhanced Laplacian Threshold technique (IELT) is applied
in TFVODT framework to perform efficient moving object
detection. In TFVODT framework, the IELT technique is
done with the help of Gaussian-based Neighbourhood
Intensity Proportion (GNIP). The architecture diagram of
Threshold Filtered Video Object Detection and Tracking
technique is shown in Fig-1.
Fig-1: Architecture of Threshold Filtered Video Object
Detection and Tracking technique

_______________________________________________________________________________________
As shown in Fig-1, TFVODT framework initially takes the
video file as input. Next, filtering techniques are applied on
the video frames for object segmentation and tracking. In
TFVODT framework, two type of filtering techniques are
used to remove the noise video frame and to improve the
video quality. The TFVODT framework performs the object
segmentation task with the help of Median Filter-based
Enhanced Laplacian Thresholding technique. With the
segmented object, then effective object tracking is
performed based on Color Histogram-based Particle Filter
that measures the likelihood function, particle posterior and
particle prior function to enhance the object tracking
accuracy. Finally, with the help of moving tracked objects,
we perform a Gaussian-based Neighbourhood Intensity
Proportion for effective moving object detection.
2.1 Median Filter-Based Enhanced Laplacian Thres
holding for Moving Object Segmentation
The design of TFVODT framework performs the object
segmentation task with the help of Median Filter-based
Enhanced Laplacian Thresholding. When an input video file
is given to the TFVODT framework, video file ‘ ’ is
partitioned into video frames ‘ ’. All the video frames
obtained do not possess characteristics such as quality,
brightness or contrast. With the intention of obtaining the
high quality video images without compromising the
quality, brightness or contrast, preprocessing has to be
performed for each input video file. Mostly several frames
in video will not be of fine quality, perfect size or with good
brightness and contrast.
An Un-symmetric Trimmed Median Filter is applied on the
video frames with the aim of reducing the Peak Signal
Noise-to Ratio and it removes the noisy frames by using a
three dimension sliding window model that is
mathematically formulated as,
(1)
From (1), the video frame ‘ ’ perform denoising using
‘ ’ sliding window. After that the median value is
evaluated to replace the noisy pixels and is formulated as
follows for a given ‘ ’ of frame ‘ ’,
(2)
Hierarchy-based Laplacian Thresholding is used in
TFVODT framework to pre-process video frames for
processing the complex video image frame by frame. The
Laplacian Thresholding performs video frame fusion with
the help of pattern selective model. This is because the
complex video image is obtained not a pix-by-pixel, but a
frame at a time. The main objective of using Laplacian
Thresholding in TFVODT framework is the decomposition
of thresholding on each video frames unite all these
decompositions to generate a complex video image.
In TFVODT framework, The Laplacian Thresholding
technique develops a thresholding for each source video
frame image by applying video frame fusion for each level
of thresholding with the help of frame selection decision
model. The frame selection decision model selects the Most
Significant Pixels (MSP) from the source video frame
images, while discarding the Least Significant Pixels (LSP).
The frame selection decision model is mathematically
formulated as given below.
(3)
From (3), the fused video frame image ‘ ’ is obtained using
‘ ’ and ‘ ’, the input video frames. This Frame selection
decision model significantly reduces the mean square error
rate in TFVODT framework. Median Filter-based Enhanced
Laplacian Thresholding technique effectively segment the
video objects in TFVODT framework which results in
improving object segmentation accuracy.
2.2 Color Histogram-Based Particle Filter For Moving
Object Tracking
After segmenting the video objects, TFVODT framework
using Color Histogram-based Particle Filter with aiming at
improving the object tracking accuracy. The objective of
using Color Histogram-based Particle Filter is to tracking of
moving objects on segmented video frames. The block
diagram of Color Histogram-based Particle Filter is shown
in Fig-2.
Fig-2: Block diagram of Color Histogram-based Particle
Filter
As shown in figure, the Color Histogram-based Particle
Filter performs three steps in TFVODT framework. The first
step evaluates the likelihood function of the color histogram
of the objects to be tracked. The second and third step
evaluates the particles posterior and prior function with
aiming at improving the object tracking accuracy.

_______________________________________________________________________________________
2.3 Improvisation of Enhanced Laplacian Threshold
For Efficient Moving Object Detection
Improvisation of Enhanced Laplacian Threshold techniques
(IELT) is applied in our TFVODT framework with the
objective of enhancing the object detection accuracy. IELT
techniques preformed with the help of Gaussian-based
Neighbourhood Intensity Proportion (GNIP). GNIP is used
for efficient moving object detection to identify the target
position in each frame by applying the Neighbourhood
Intensity Proportion. The feature vectors are mined with the
help of correlation in object detection phase. The TFVODT
framework deals with a neighbourhood intensity proportion
through incorporating the sophisticated object detection
algorithm such as Gaussian Contour Foreground Video
Detection. The algorithmic process of Gaussian Contour
Foreground Video Detection is shown in Fig-3.
Input: Video ‘ ’, Video Frame
‘ ’, weight of contour video
object ‘ ’, weight of foreground video object
‘ ’, number of foreground pixels in contour
video object ‘ ’, number of contour pixels in
foreground video objects ‘ ’
Output: Enhanced video object detection accuracy
Step 1: Begin
Step 2: For each video frame ‘ ’
Step 3: Evaluate neighbourhood intensity proportion
Step 4: Measure the actual value of pixels
Step 5: Evaluate Neighbourhood Intensity Proportion
Distribution
Step 6: Evaluate Enhanced Moving Video Object
Contour
Step 7: Measure correlation coefficient
Step 8: End for
Step 9: End
Fig-3: The algorithmic of Gaussian Contour Foreground
Video Detection
From the above figure, the Gaussian Contour Foreground
Video Detection algorithm process is described in five steps.
For each video frame, the first step evaluates the
neighborhood intensity proportion to obtain the foreground
video object. After that, the second step measures the actual
value of pixels for removing the noise present in the video
object. After removing the noise, the third step evaluates the
Neighborhood Intensity Proportion Distribution to identify
the foreground video object. The fourth step evaluates the
Enhanced Moving Video Object Contour. Finally, the
correlation coefficient is calculated with the objective of
improving the moving object detection accuracy.
3. EXPERIMENTAL SETUP
The Threshold Filtered Video Object Detection and
Tracking (TFVODT) framework conduct experimental work
on MATLAB. The TFVODT framework uses the video files
and their sizes are listed in Table-1 for moving object
detection. The video files are obtained from Internet Archive
501(c) (3), a non-profit organization. The Internet Archive
contains texts, audio, moving images, and software as well
as archived web pages.
The video file information listed in table 1 contains the
name of the video file, resolution of the video files and their
size respectively for evaluating the TFVODT framework.
The video used for moving object detection and tracking
using TFVODT framework is shown below with detailed
information. Experimental evaluation using TFVODT
framework is conducted on various factors such as object
segmentation accuracy, Peak Signal to Noise Ratio, object
tracking accuracy, Mean Square Error and object detection
accuracy with respect to different videos and video frames.
Table-1: Video file information
Name
Video file information
Video
frames
Resolution Size (KB)
Blossom.avi 216 * 192 349.5
Sample.avi 256 * 240 113.6
Vehicle.avi 510 * 420 323.7
Atheltic.avi 854 * 480 905.3
Person.avi 320 * 240 936.2
Flower.avi 350 * 240 454.5
Rose.avi 458 * 213 635.2
4. DISCUSSION
The performance of TFVODT framework is compared with
exiting two methods namely, Self Crossing Detection for
Parametric Active Contours (SCD-PAC) [1], Automatic
Estimation of Multiple Motion (AEMM) [2]. The
performance of TFVODT framework is evaluated along
with the following metrics.
4.1 Impact of Object Segmentation Accuracy
The object segmentation accuracy is defined as the ratio of
objects being segmented to the total number of
frames/second. The object segmentation accuracy is
formulated as below

_______________________________________________________________________________________
(4)
From (4), ‘ ’ represents the object segmentation
accuracy whereas ‘ ’ refers to the objects being correctly
segmented. When the object segmentation accuracy is
higher, the method is said to be more efficient and it is
measured in terms of percentage (%).
Table-2: Tabulation for object segmentation accuracy
Number of
frames/second
Object segmentation accuracy (%)
TFVODT
SCD-
PAC
AEMM
10 75.25 66.35 61.54
20 79.54 71.25 66.41
30 84.47 75.68 70.57
40 87.87 81.87 74.98
50 89.17 84.69 78.94
60 93.98 87.1 82.19
70 95.47 91.81 86.87
The object segmentation accuracy of TFVODT framework
is elaborated in Table-2. We consider the framework with
different number of frames/second in the range of 10 frames
to 70 frames for experimental purpose using MATLAB. The
performance of proposed TFVODT framework is compared
with existing two methods namely, SCD-PAC [1], AEMM
[2].
Fig-4: Measure of object segmentation accuracy
Fig-4 shows the impact of object segmentation accuracy
with respect to varying frames being sent per second in the
range of 10 to 70. As illustrated in Figure, the proposed
TFVODT framework performs relatively well when
compared to two other methods SCD-PAC, AEMM. This is
because of the application of Median Filter-based Enhanced
Laplacian Thresholding that efficiently removes the noise
frame for improving the video quality. Therefore, the object
segmentation accuracy using TFVODT framework is
improved by 7.95% as compared to SCD-PAC [1] and
14.12% as compared to AEMM [2] respectively.
4.2 Impact of Peak Signal To Noise Ratio
Peak Signal-to-Noise Ratio calculates the ratio between the
reference video frame and the distorted video frame being
detected in a video file. The higher the PSNR, the closer the
distorted video frame is to the original. Therefore, higher
PSNR value associate with higher quality image (i.e.
detected image) and is mathematically formulated as given
below.
(5)
(6)
From (5), the mean square error ‘ ’ is the difference
between the actual frame size ‘ ’ and the estimated frame
size ‘ ’ being detected. From (6), the peak signal-to-noise
ratio ‘ ’ is evaluated using the unsigned integer data
type (with size 255) with respect to mean square error rate
‘ ’ respectively.
Table-3: Tabulation for PSNR
Size of
video file
(KB)
Peak Signal to Noise Ratio (%)
TFVODT SCD-PAC AEMM
113.6 26.65 20.35 16.32
323.7 31.54 25.13 20.54
349.5 37.54 30.03 24.65
454.5 39.87 32.19 28.98
635.2 42.51 35.43 32.14
905.3 44.65 38.15 36.11
936.2 47.87 41.82 40.21
To determine the performance of the PSNR for moving
object detection, comparison is made with two other existing
methods SCD-PAC [1], AEMM [2]. In Fig-5, the size of
video frames is varied between 100 KB and 1000 KB. From
the figure it is illustrative that the PSNR using the proposed
TFVODT framework is reduced when compared to existing
methods.
Fig-5: Measure of Peak Signal to Noise Ratio
The PSNR rate for moving object detection is reduced by
applying Color Histogram-based Particle Filter in TFVODT
framework. With the application of Color Histogram-based
Particle Filter, posterior function and prior function is
evaluated for each video frames in an iterative manner that
provides the results with respect to varying frame size which
in turn reducing the PSNR by 18.15% as compared to SCD-
PAC [1]. In addition, by applying the state space

_______________________________________________________________________________________
representation model between the state space and current
measurement, by formulating an initial posterior function
minimizes the PSNR for moving object detection by 27.83%
as compared to AEMM [2].
4.3 Impact of Object Tracking Accuracy
The object tracking accuracy is defined as the ratio of
objects being tracked to the total number of frame / second.
The object tracking accuracy is formulated as below
(7)
From (7), ‘ ’ signifies the object tracking accuracy
whereas ‘ ’ refers to the objects being correctly tracked.
When the object tracking accuracy is higher, the method is
said to be more efficient and it is measured in terms of
percentage (%).
Table-4: Tabulation for object tracking accuracy
Number of
frames/second
Object tracking accuracy (%)
TFVODT
SCD-
PAC
AEMM
10 77.22 74.83 71.24
20 81.65 78.21 75.64
30 86.25 83.31 80.87
40 79.73 76.29 73.69
50 84.89 81.16 79.34
60 87.32 84.89 81.27
70 90.78 87.32 85.11
The object tracking accuracy for multiple moving object
tracking using TFVODT, SCD-PAC, and AEMM is
elaborated in Table-4. We consider the framework with
different number of frames/second in the range of 10 frames
to 70 frames for experimental purpose using MATLAB.
In Fig-6, the object tracking accuracy using the proposed
TFVODT framework is higher when compared to two other
existing methods SCD-PAC [1], AEMM [2]. With respect to
the increasing number of videos and frames/second, the
object tracking accuracy is increased though not observed to
be linear (i.e. due to the presence of noise), but shows
gradual improvement by using the Color Histogram-based
Particle Filter when compared to SCD-PAC, AEMM.
Fig-6: Measure of Object tracking accuracy
Fig-6 illustrates the rate of object tracking accuracy and is
observed to be higher using TFVODT framework. The rate
of object tracking accuracy is verified for different number
of frames. By applying Color Histogram-based Particle
Filter, the object tracking accuracy is improved using
TFVODT framework by 4.80% as compared to SCD-PAC
[1]. In addition, by evaluating the likelihood function of the
color histogram in TFVODT framework, the object tracking
accuracy is improved by 7.98% as compared to AEMM [2].
4.4 Impact of Mean Square Error
The Mean Square Error (MSE) in TFVODT framework is
used to measure the video quality being obtained during
multiple object tracking. The MSE represents the error
between the tracked frames and the original video frame.
When lower the value of MSE, the method is said to be
more efficient and it is measured in terms of db.
(8)
From (8), the mean square error ‘ ’ is the difference
between the actual frame size ‘ ’ and the estimated frame
size ‘ ’ being tracked.
Table-5: Tabulation for mean square error
Size of video
frame (KB)
Mean Square Error (db)
TFVODT
SCD-
PAC
AEMM
113.6 21.63 25.34 29.24
323.7 23.87 26.44 30.87
349.5 25.98 30.22 35.47
454.5 30.21 32.77 37.59
635.2 37.34 39.35 44.88
905.3 43.87 46.25 49.14
936.2 51.57 55.35 57.68
Table-5 represents the mean square error rate with respect to
different size of video frame using MATLAB and
comparison is made with two other methods, namely SCD-
PAC [1], AEMM [2].
Fig-7: Measure of Mean Square Error
Fig-7 shows the mean square error rate with size of the
video frame taken as the input for moving object detection.
As illustrated in Figure, the proposed TFVODT framework
provides minimum mean square error rate when compared

_______________________________________________________________________________________
to two other methods SCD-PAC, AEMM. This is because of
the application of Median Filter-based Enhanced Laplacian
Thresholding technique in TFVODT framework. With the
application of Median Filter-based Enhanced Laplacian
Thresholding, Bayes estimation of posterior and prior
function is evaluated in an efficient manner. This in turn
helps in reducing the mean square error for moving object
detection using TFVODT framework by 10.12% as
compared to SCD-PAC [1]. In addition, the TFVODT
framework using Bayes Sequential Estimation that takes
state space as input helps in decreasing the mean square
error by 24.22% as compared to AEMM [2].
4.5 Impact of Object Detection Accuracy
The object detection accuracy is defined as the ratio of
objects being detected to the total number of frames/second.
The object detection accuracy is mathematically formulated
as below
(9)
From (9), ‘ ’ signifies the object detection accuracy
whereas ‘ ’ refers to the objects being correctly detected.
When the object detection accuracy is higher, the method is
said to be more efficient and it is measured in terms of
percentage (%).
Table-6: Tabulation for object detection accuracy
Methods
Object detection accuracy
(%)
TFVODT 88.35
SCD-PAC 75.12
AEMM 69.31
The object detection accuracy for TFVODT framework is
elaborated in Table-6 and comparison made with two other
methods SCD-PAC, AEMM respectively. We consider the
method with seven video objects and different number of
frames/second in the range of 10 frames to 70 frames for
experimental purpose using MATLAB.
Fig-8: Measure of object detection accuracy
From the Fig-8, it is illustrative that the object detection
accuracy using the Improvisation of Enhanced Laplacian
Threshold techniques in TFVODT framework is higher than
when compared to the existing methods SCD-PAC, AEMM
method respectively. This is because of the application of
Gaussian mixture model with neighbourhood intensity
proportion in TFVODT framework that measures the
correlation coefficient of the moving video object in an
efficient manner using Gaussian Contour Foreground Video
Detection algorithm with Neighbourhood Intensity
Proportion Distribution. Therefore, the object detection
accuracy is improved by 14.97% when compared to SCD-
PAC [1] method and 21.55% improvement when compared
to AEMM [2] respectively.
5. CONCLUSION
In this work, an effective novel framework is designed
called as Threshold Filtered Video Object Detection and
Tracking (TFVODT) for effective detection of moving
objects in video frames. The TFVODT technique improves
the object segmentation accuracy with reduced PSNR by
applying Median Filter-based Enhanced Laplacian
Thresholding technique. The goal of TFVODT technique for
video surveillance is to improve the object segmentation
accuracy, object tracking accuracy and object detection
accuracy by demonstrating with training and test video
objects obtained from Internet Archive 501(c) (3), a non-
profit organization which significantly contributes to the
relevance. With the experiments conducted for TFVODT
technique, it is observed that of video object detection for
different video samples provided more accurate results as
compared to existing detection methods. The results show
that TFVODT technique provides better performance with
an improvement of object detection accuracy by 18% and
reduces the PSNR rate by 23% when compared to state of
the art works.
6. REFERENCES
[1]. Arie Nakhmani, Allen Tannenbaum, “Self-Crossing
Detection and Location for Parametric Active
Contours”, IEEE Transactions on Image Processing,
Volume 21, Issue 7, July 2012, Pages 3150 – 3156.
[2]. Manya V. Afonso, Jacinto C. Nascimento and Jorge S.
Marques, “Automatic Estimation of Multiple Motion
Fields From Video Sequences Using a Region
Matching Based Approach”, IEEE Transactions on
Multimedia, Volume 16, Issue: 1, January 2014, Pages
1-14.
[3]. Taehee Lee , Stefano Soatto, “Video-based descriptors
for object recognition”, Elsevier, Image and Vision
Computing, Volume 29, Issue 10, September 2011,
Pages 639–652.
[4]. Shih-Wei Sun, Yu-Chiang Frank Wang , Fay Huang ,
Hong-Yuan Mark Liao, “Moving foreground object
detection via robust SIFT trajectories”, Journal of
Visual Communication and Image Representation,
Volume 24, Issue 3, April 2013, Pages 232–243.
[5]. Hamid Izadinia, Imran Saleemi, and Mubarak Shah,
“Multimodal Analysis for Identification and
Segmentation of Moving-Sounding Objects”, IEEE
Transactions on Multimedia, Volume 15, Issue 2,

_______________________________________________________________________________________
February 2013, Pages 378-390.
[6]. Subhabrata Bhattacharya, Haroon Idrees, Imran
Saleemi, Saad Ali and Mubarak Shah, “Moving Object
Detection and Tracking in Forward Looking Infra-Red
Aerial Imagery”, Springer, Machine Vision Beyond
Visible Spectrum, May 2011, Pages 1-32.
[7]. Manya V. Afonso, Jacinto C. Nascimento and Jorge S.
Marques, “Automatic Estimation of Multiple Motion
Fields From Video Sequences Using a Region
Matching Based Approach”, IEEE Transactions on
Multimedia, Volume 16, Issue: 1, January 2014, Pages
1-14.
[8]. Henry Medeiros, Germán Holguín, Paul J. Shin,
Johnny Park, “A parallel histogram-based particle filter
for object tracking on SIMD-based smart cameras”,
Elsevier, Volume 114, Issue 11, November 2010, Pages
1264–1272.
[9]. Tom Mélange, Mike Nachtegael, and Etienne E. Kerre,
“Fuzzy Random Impulse Noise Removal From Color
Image Sequences”, IEEE Transactions on Image
Processing, Volume 20, Issue 4, April 2011, Pages
959-970.
[10].Samuele Salti, Andrea Cavallaro, Luigi Di Stefano,
“Adaptive Appearance Modeling for Video Tracking:
Survey and Evaluation”, IEEE Transactions on Image
Processing, Volume 21, Issue 10, October 2012, Pages
4334-4348.
[11].Pojala Chiranjeevi and Somnath Sengupta,
“Neighborhood Supported Model Level Fuzzy
Aggregation for Moving Object Segmentation”, IEEE
Transactions on Image Processing, Volume 23, Issue 2,
February 2014, Pages 645 – 657.
[12].Barga Deori and Dalton Meitei Thounaojam, “A
Survey on Moving Object Tracking in Video”,
International Journal on Information Theory (IJIT),
Vol.3, No.3, July 2014, Pages 31 – 46.
[13].Sukanyathara J and Alphonsa Kuriakose, “An
Optimized Framework for Detection and Tracking of
Video Objects in Challenging Backgrounds”, The
International Journal of Multimedia & Its Applications
(IJMA), Vol.6, No.4, August 2014, Pages 27 – 38.
[14].Daniela Giordano, Francesca Murabito, Simone
Palazzo, Concetto Spampinato and University of
Catania, “Superpixel-based Video Object Segmentation
using Perceptual Organization and Location Prior”, The
IEEE Conference on Computer Vision and Pattern
Recognition (CVPR), 2015, Pages 4814-4822.
[15].Zhengyang Wu, Fuxin Li, Rahul Sukthankar and James
M. Rehg, “Robust Video Segment Proposals with
Painless Occlusion Handling”, The IEEE Conference
on Computer Vision and Pattern Recognition (CVPR),
Computer Vision Foundation, 2015, Pages 4194 –
4203.
[16].Jianping Han, Tian Tan, Longfei Chen and Daxing
Zhang, “Moving Objects Representation for Object
Based Surveillance Video Retrieval System”,
International Journal of Security and Its Applications,
Vol.8, No.2, March 2014, Pages 315 – 322.
[17].Safvan Vahora, Narendra Chauhan, Nilesh Prajapati,
“A Robust Method for Moving Object Detection Using
Modified Statistical Mean Method”, International
Journal of Advanced Information Technology (IJAIT),
Vol. 2, No.1, February 2012, Pages 65-73.
[18].Ms Jyoti J. Jadhav, “Moving Object Detection and
Tracking for Video Survelliance”, International Journal
of Engineering Research and General Science Vol. 2,
No. 4, July 2014, Pages 372-378.
[19].Mrinali M. Bhajibhakare, Pradeep K. Deshmukh, “To
Detect and Track Moving Object for Surveillance
System”, International Journal of Innovative Research
in Computer and Communication Engineering, Vol. 1,
No. 4, June 2013, Pages 945-949.
[20].R. Kwitt1, N. Vasconcelosb, S. Razzaquec, S.
Aylward, “Localizing Target Structures in Ultrasound
Video -A Phantom Study”, Elsevier, Volume 17, Issue
7, October 2013, Pages 712–722.
[21].P.Vijayakumar and A.V.Senthil Kumar “Moving
Object Segmentation using Enhanced Laplacian
Thresholding Method for Video
Surveillance”,IJCA,Volume 100,Issue 4,August
2014,Pages 13 -17.
[22].P.Vijayakumar and A.V.Senthil Kumar “A Novel
Framework for Specific Object Tracking in the
Multiple Moving Objects
Environment”,IJAER,Volume 10, Number 16 (2015)
Pages 36540 - 36545.
[23].P.Vijayakumar and A.V.Senthil Kumar “Moving
Object Segmentation Using Median Filter
[24].-Based Enhanced Laplacian Thresholding
“EJSR,Volume 133,Issue Pages 415- 427.
[25].P.Vijayakumar and A.V.Senthil Kumar” Improvising
Enhanced Laplacian Thresholding Technique for
Efficient Moving Object Detection in Video
Surveillance” IOSRJEN, Vol. 06, Issue 01 (January.
2016), ||V1|| PP 42-52.

Threshold based filtering technique for efficient moving object detection and tracking in video surveillance

More Related Content

What's hot (19)

Viewers also liked (10)

Similar to Threshold based filtering technique for efficient moving object detection and tracking in video surveillance (20)

More from eSAT Journals (20)

Recently uploaded (20)

Threshold based filtering technique for efficient moving object detection and tracking in video surveillance