International Journal of Emerging Trends & Technology in Computer Science (IJETTCS)

Web Site: www.ijettcs.org Email: [email protected]

Volume 5, Issue 2, March - April 2016

ISSN 2278-6856

Survey on Serial and Parallel Computing

approaches for Real-Time Object Detection,
Recognition and Tracking Systems
K.Sivakumar1 , Dr. P.Shanmugapriya2
1

Research Scholar, Department of CSE SCSVMV University, Kanchipuram, India

Associate Professor, Department of IT SCSVMV University, Kanchipuram, India

Abstract
Object Detection, Recognition and Tracking operations are
the important operations involved in machine vision and
computer vision applications such as video surveillance,
intelligent vehicle systems, person identification system,
industrial product quality and quantity checking and many
more. In the real time systems, it requires high speed
computing approaches and devices to perform these
operations. A parallel computing approach is the recent
approach after the development of multi-core processor.
Parallel computing method gives high performance result
then the traditional way of serial computing. These computing
approaches can be developed in two ways; one is software and
another is hardware approach. There is a need for explore all
these methods and approaches collectively to taken in to
consideration for the efficient implementation of the real-time
applications with high accuracy and good result. This paper
presents the survey on Real-time object detection, recognition
and tracking systems for computer vision and machine vision
systems through serial, parallel, hardware and software
approaches and also it presents the different algorithms,
methods, result analysis, limitations evaluation metrics and
the best approaches for future research work.

Keywords: objects; detection; tracking; recognition;

hardware; software, serial; parallel, FPGA.

1. INTRODUCTION
Machine vision and Computer vision are the two broad
area where the image and video analytic operations being
carried out. Machine vision is the technology is used to
replace or complement manual inspection and
measurement with digital camera and image processing
[2]. It is the distinct field which provides imaging based
automatic inspection, analysis, process controls, robotic
guidance, and industrial product quality control and more
[1][3]. Computer vision is another popular area where the
image/video capturing, refining, processing, analyzing,
understanding the images takes place. In both of these
area; Object detection, recognition and tracking are the
three important operations involved for many of the real
time machine vision and computer applications [4]. At
first, Object detection involves the extractions of
information from the image or video, processing of the
information and determining whether the object is present
in the image or video frames or not [4]. Object detection
operation is the important and base operation to perform
other operations. Secondly, Object recognition is the
process of recognition of known object and unknown

Volume 5, Issue 2, March April 2016

objects is present in the image or video frames [5]. This

operations mainly used for searching or recognition [5] of
the particular object is present in the image or video
frames. Finally, object tracking is the process of tracking
the moment of the objects or to track the path of the
objects in the video frame [6]. To perform these operations
in real time machine vision and computer vision systems
applications we need of high performance computing
approach [4]; it may be hardware or software approach.
There are many algorithms; methods are proposed for
object detection, tracking and recognition [6][7], each
approaches were implemented and tested with hardware
and software. The following are the methods for object
detection a). Frame difference, b). Optical flow, c).
Background subtraction and many more [6][7], and then
for object recognition methods are shape based, motion
based, color based, texture based and more; finally object
tracking is concern point based, kernel based and
silhouette-based and more[6][7]. The above mentioned
methods are implemented with hardware as well as
software approach and achieve considerable result, but all
these are consider only little bit care on to do power
optimization, parallelization, accuracy and time
consuming. There are some challenges need to face for
successfully implementation of the object detection,
tracking and recognition operations. To detect faces in in
any image and video the following challenges like
illuminant variations, dynamic moving scene, and others
need to be consider while we implementing. For object
tracking using faces we need to tackle the following
challenges like dynamic moving objects, scaling of the
objects and others; and finally for object recognition faces
occlusion, image or video sizes, object orientation[] etc are
the challenges.
This paper presents the various techniques, methods,
algorithms which were used for object detection, tracking
and recognition operations; and also discuss about the
limitations, advantages and disadvantages of the various
methods, future works, enhancements, performance
evaluations, comparisons charts between the various
methods. This paper is organized with many sections; in
Section I, discuss about the hardware based object
detection, software based object detection in parallel. In
section II discuss about the object tracking through
hardware as well as software approaches in both parallel
Page 87

International Journal of Emerging Trends & Technology in Computer Science (IJETTCS)

Web Site: www.ijettcs.org Email: [email protected]
Volume 5, Issue 2, March - April 2016
and serial computing approach. In section III discuss
about the object recognition method in hardware and
software approach in parallel or serial. In section IV
discuss about the comparisons with charts between the
various methods, what we have discussed, finally the
conclusion and future work being discussed.

2. OBJECT DETECTION RECOGNITION

AND TRACKING CLASSIFICATION
This section discusses about the broad classification for
object detection, tracking and recognition [31][32] and its

ISSN 2278-6856

sub classification. It is important to study and understand

the various classification models to detect, track and
recognition to choose which method is suitable for what
and when. If we know these classifications clearly then
only we can implement right application using right
methods. Let us take an example of object detection
[31][32]; background subtraction can be used for many
applications implementation. Here we list out the various
types and sub types for further analysis.

Video / Image operations

Object detection

Object tracking

Background subtraction

Object
recognition/Classification

Region based
Shape based

Recursive, Non
Recursive

Feature based
Color, Shape

Feature based

Motion Based

Point tracking
Color and shape

Color based
Recursive, Non
Recursive

Motion based

Texture based
Kernel based

Frame differencing
Optical flow
Gaussian Mixture

Optical flow

Template based

Silhouette

Fixed template
Deformable
Template

Texture Based

Fig. 1 Object detection, tracking and recognition classification

There are some advanced methods for object recognition, tracking and recognition were introduced. These methods are
also listed in literature survey.
3.1.1). AdaBoost Based Object Detection
In[4] their work, they propoed and implemented generic,
3. LITERATURE SURVEY
flexible parallel adaBoost based architecture for efficient
3.1). Object detection, recognition and tracking in
object detection method. FPGA, and xilinx are the device
hardware based serial or parallel computing
and software used for this experiments. It was designed
approach: the following were the related papers which
for detction several objects from up to 1024x768 size of
deals about the object detection tracking and recognition
images, and for video processing is concern it perform
in hardware based parallel computing approaches.
64-139 frame per seconds. This proposed work used
systoloc arrays of computarion to perform integral image

Volume 5, Issue 2, March April 2016

Page 88

International Journal of Emerging Trends & Technology in Computer Science (IJETTCS)

Web Site: www.ijettcs.org Email: [email protected]
Volume 5, Issue 2, March - April 2016
calculation, and for up scaling of images and down
scaling of the image can be done using image pramid
concept. This parallel archicture was evaluated by using
xlinx Vertex II pro FPGA. It is suitable for several type of
budject and applications. Power optimization is left to
future development. Fig. 3 shows the detection result of
different applications for various sizes for image, feature
sizes, frames and various objects.
3.1.2). Cascade Support Vector Machines
In [22], It proposed hardware architecture for acceleration
of Cascade SVM used to desing intelligent embedded
systems for online real-time calssification applications. It
reduces the hardware resources and implemented using
Spartan-6 eld-programmable gate array (FPGA)
platform and evaluated for object detection on 800 600
(Super Video Graphics Array) resolution images. It was
boosted by nueral network to achives 40 frames per
seconds. It achieves face detection ranges between 78%
and 80%.
3.1.3). Support Vector Machines for 3D object
recognition
Massimiliano pointil et al [33], in this proposed method,
given a set of points which belong to either of two classes.
One is linear and another is a non linear based on the
hyperplane. Here they used linear SVM methos for 3D
object recognition methods. It used 7200 images for 100
different objects were detected. It does not require feature
exatraction and performance recognition on images
regard as point of space of high dimention without
estimations pose. It is well suited for aspect based
recognition.
3.1.4). Adaboost algorithm with Modified Census
Transform method
Dongkyun Kim et al[35], they proposed eye detection
method based on the hardware. They developed a method
for eye detection for various applications based on the
adaBoost algorithm with modified census transform
methods to detect eye on the images only not in the video
sequences. It is parallelize hardware architecture method
were introduced for eye detection with accuracy and
downscaling the images are also done. it was
implemented using FPGA and Verilog HDL. This method
can detect eye with 0.15 ms with maximum estimated
frequency of 106 MHz with expected 240 fps in VGA
image. Here the detection rate achieved 93% however the
detection rate may drop in glasses.
3.2). Object detection, recognition and tracking in
software based serial or parallel computing or GPU
approach
The following were the related papers which deals about
the object detection, recognition and tracking in software
Tracking feedback and CA should be feasible and flexible
with parameters of different PTZ cameras. Need high
speed background modeling algorithm for different
tracking methods. This work was developed in serial
computing platform using c++ and opencv. PTZ kind of
object detection and tracking methods need high speed
computing platform, this work left to the future

Volume 5, Issue 2, March April 2016

ISSN 2278-6856

based serial, parallel computing or GPU computing

approaches.
3.2.1). Fuzzy Color Difference Histogram and Based
Background Subtraction
In [23] Deepak Kumar Panda et al., The proposed works
demonstrated on publically available various challenges
video sequences such as Campus, Waving Trees, Water
Surface, Curtain, Lobby and PETS2 006. The test
sequences can be down- presented their work for
detection of moving object with complex scene such as
dynamic background, illumination variation, and
camouage. It proposed robust background subtraction
technique with three important methods, firstly the Color
Difference Histogram(CDH), it reducess no of false
errors. Secondly, color difference is fuzzified with
gaussion membership function. Finally, novel fuzzy color
difference histogram, it reducess the large computation
and also lession the intensity variation[23] it fails to
detect shadow cast by the object and gave considerable
performance. Hence, even though they used in industrial
automotion, traffic monitoring and other real time
systems need high speed computing, this draback leed to
the development of the future work in parallel or GPU
computing approach.
In [29], this approach is used to automatically tracking
and detecting unusual suspecious moments in CCTV.
This works based on the optain foreground object by
subtracting background objects. It is classified people and
inanimated objects, tracked realtime blob matching
techniques and all object activities are classified using
semantic based-approach. This approach gives real-time
performance, adaptability, robustness against clutter,
camera nonlinearities and eliminating machine learning
based training methods.
In [26], new approach for video survilance object
recognition algorithm were presented, based on shape
feature, color histogram of the objects. The output for the
recognition insensitive to the illuninamce, shadow,
rotation and distance. This method adapted adaboost
method and it gave considerable result.
In [27] Ning Liu, Hefeng Wu, and Liang Lin, they
presented hierarchical background model, frame
registration and object tracking are done in serial. There
are some issues in the proposed work, firstly need for
good automatic algorithm for different scenes. Secondly,

development of high speed background modeling in

software baser parallel or GPU based or parallel hardware
based approach [27].

Page 89

International Journal of Emerging Trends & Technology in Computer Science (IJETTCS)

Web Site: www.ijettcs.org Email: [email protected]
Volume 5, Issue 2, March - April 2016
3.2.2). Region of Interest and spacial and temporal
exatraction
In [28], this work propose effective traffic survilance
system for traffic vehicle classifying and counting at night
time environments. This proposed work uses
segmentation process based on the adaptive threshold to
extract bright objects, then this will feed in to the spacial
and temporal features fot the vehicle pattern extraction. It
performed tracking and counting for vehicle video with
320x180 pixels size only. The vehicle type classification
function can be further improved and extended by
integrating some sophisticated machine learning
techniques such as SVM and also further classified into
such as buses, trucks, and light and heavy motorbikes etc
3.2.3). Viola-Jones casecade classifiers
In [27], this work proposed object detection and
recognition task done using Viola-Jones casecade
classifiers. This work was implemented and parallel
methodologies for integral image calculation, windows
processing. Compared with the tradidinal CPU program
and this approach achieves 22.42 times speed up for

ISSN 2278-6856

detection pahses and 1668.56 times speed up for

recognition phase using NVIDA CUDA GTX 570
graphics card. It trained up to 2000 images testing 40
images with intel(i7 Processor) CPU. It gives good real
time performance. This proposed work left to
implementation of recognition using high end NVIDA
kepler architecture and introduction new classified to
perform detection and recognition more and more faster
for the future development.
3.2.4) Histogram of Gradient and SVM
In [30], GPU based real time detection and recognition
for 48 different signs. Histogram of Gradient method with
optimized parallel algorithm using GPU system was used
to detect the sign and to recognize, SVM was used. There
were 32 scale level of processing recognition, for more
accurate recognition SVM method was used. It achieves
27.9 fps with 1628x1236 resolution, detection rate is
91.69% and for recognition rate is about 93.77%. for
future work,there is need for use of previous frame can be
used to increase the performance.

[4] it usesd
AdaBoost based
object detection
Algorithm, parallel
systolic
computation

64-139 fps, up
to 1024x 768
image size,
Xilinx Virtex
II Pro FPGA

[23] Background
subtraction with
color difference
histogram (CDH),
color difference is
fuzzied with a
Gaussian
membership, fuzzy
color
difference
histogram (FCDH)

MATLAB, i5,
3.10 GHz,
4GB RAM,

[22], Cascade
Support Vector
machines, face
detection

40 frames/s
800x600(SVG
A)
Spartan-6
eldprogrammable
gate array
(FPGA)

Flexible
architecture
used several
types of
applications

Power
optimization,
system level
optimization,
systolic array
sizes, need
development of
processor for
online training,
dynamic and
autonomous
environments and
situations

CDH reduces
the number
of false errors,
it fails to detect
shadow cast by
the object,
industrial
automation,
trafc
monitoring,
security and
surveillance

Need parallel
processing to
achieve
performance

Used to desing
online realtime
classification
Low cost
parallel
processor,
tackle large
scale problesm

Need to design
>80 % for higher
resoluton images

Volume 5, Issue 2, March April 2016

Execution
samples

Execution
Result

Future
works

Advantages
and
Disadvantag
es

Computation
al Time/
Frames per
seconds,
Image size,
Tools used

Methods,
Concepts
/Algorithm
Used

TABLE-1.Sample Object Detection, Tracking, Recognition Methods, Experimental Result and Analysis

in [4]
In [4]
Average computation time per pixels in
clocks

in [23]

In [22]

In [23]

In [22]

Page 90

International Journal of Emerging Trends & Technology in Computer Science (IJETTCS)

[30] Window-based
Histogram
of
Gradient, SVM to
classify sign

27.9 fps, pixels

of 1,628
1,236
resolution,
BelgiumTS
Dataset,
for
detection
91.69 %, for
recognition
93.77%
OpenCV-GPU
based
approach

capable
of
detecting
and
recognizing 48
classes
of
trafc signs in
any sizes on
each
image
frame,
no
information
from previous
frame
are
required, less
detection and
recognition
accuracy

Replace serial
computation to
parallel to get fast
counting and
classifying
vehicle at night
time

Cars
Motorbike
s
Time span
of the
video

99.8
3
91.3
0

98.1
3
90.5

99

50
min

50
min

50
min

90.9
0

[28]

To
increase
detection
and
recognition
accuracy
and
performance.
Need hardware
based architecture
for real time
system

4.CONCLUSION AND FUTURE DIRECTION

This paper provided comprehensive survey of research on
object detection, tracking and recognition categories for
various approaches and also categorized the various general
classifications for object detection, tracking and recognition
based on serial, parallel computation, hardware and
software based approach. With development of parallel
computing it differ from sequential computing like GPU
method. There has been tremendous changes have been
made in object detection, tracking and recognition, but this
method is suitable for limited size of memory can be used.
Now a day all the computer was developed in multi core
processor system this will increase and lead the parallel
computing techniques. There also need for memory
reconfigurable methods to do the above operations on the
real-time system. The hardware based, and reconfigurable
architecture is best suited for object detection, tracking, and
recognition. There are many modern reconfigurable
hardware description languages were introduced to do the
video and image to design hardware architecture in parallel.
There are number of scope for implementing image and
video based reconfigurable hardware design for computer
vision and machine vision applications.

[4].

REFERENCES

[8].

[1]. Machine Vision, Ramesh Jain, Rangachar Kasturi,

Brian G. Schunck Published by McGraw-Hill, Inc.,
ISBN 0-07-032018-7, 1995
[2]. Machine vision introduction, SICK IVP Version 2.2,
December 2006
[3]. By Bruce G Batchelor coming to terms with machine
vision
and
Computer
vision,

[9].

Volume 5, Issue 2, March April 2016

ISSN 2278-6856

FMeasure
%

Used
for
trackig
and
counting
moving vehicle
night time.
Perform only
320x180 pixel
size,
97%
achieved

Recall
%

30 frames/ s,
dual core 2.4
GHz, 320x180
pixels size
C++ used for
implementatio
n
50 min video

Precisio
n%

[28], Region of
interest
based
object detection and
tracking for moving
vehicle,
Serial
computing
approach,
segmentation
process based on
adaptive threshold,
spatial
and
temporal features
for identify and
classify

Vahicle
s

Web Site: www.ijettcs.org Email: [email protected]

Volume 5, Issue 2, March - April 2016

[5].

[6].

[7].

http://ece631web.groups.et.byu.net/References
/Machine- Computer%20Vision.pdf
Christos Kyrkou, Student Member, IEEE, and
Theocharis Theocharides, Member, IEEE, A Flexible
Parallel Hardware Architecture for AdaBoost-Based
Real-Time Object Detection Christos Kyrkou, Student
Member, IEEE, and Theocharis Theocharides, Member,
IEEE IEEE TRANSACTIONS ON VERY LARGE
SCALE INTEGRATION (VLSI) SYSTEMS, VOL. 19,
NO. 6, JUNE 2011
Hsuan Lee, Chih-Yin Liu, Implementation of Real-time
Object Recognition System for Home-Service Robot by
Integrating , 2014 IEEE International Conference on
System Science ang Engineering (ICSSE)
Himani S. Parekh, , Darshak G. Thakore , Udesang K.
Jaliya A Survey on Object Detection and Tracking
Methods, International Journal of Innovative Research
in Computer 2014
Mrs J. Komala Lakshmi, , and Dr.M.Punithavalli, A
Survey on Performance Evaluation of Object Detection
Techniques in Digital Image Processing , IJCSI
International Journal of Computer Science Issues, Vol.
7, Issue 6, November 2010
Guo Wei, Qi Qingwen and Jiang Lili, Zhang Ping, A
NEW METHOD OF SAR IMAGE TARGET
RECOGNITION
BASED
ON
ADABOOST
ALGORITHM 2008 IEEE
Junguk Cho, Bridget Benson, Shahnam Mirzaei, Ryan
Kastner, Parallelized Architecture of Multiple
Classifiers for Face Detection, 2009 20th IEEE
International Conference on Application-specific
Systems, Architectures and Processors
Page 91

International Journal of Emerging Trends & Technology in Computer Science (IJETTCS)

Web Site: www.ijettcs.org Email: [email protected]
Volume 5, Issue 2, March - April 2016
[10]. Masayuki Hiromoto, Student Member, IEEE, Hiroki
Sugano, Student
Member, IEEE, and Ryusuke
Miyamoto,
Member,
IEEE
Partially Parallel
Architecture for AdaBoost-Based Detection With HaarLike Features IEEE TRANSACTIONS ON CIRCUITS
AND SYSTEMS FOR VIDEO TECHNOLOGY, VOL.
19, NO. 1, JANUARY 2009
[11]. Christos Kyrkou, Student Member, IEEE, and
Theocharis Theocharides, Member, IEEE A Flexible
Parallel Hardware Architecture for AdaBoost-Based
Real-Time Object Detection, IEEE TRANSACTIONS
ON VERY LARGE SCALE INTEGRATION (VLSI)
SYSTEMS, VOL. 19, NO. 6, JUNE 2011 Kalyan
Kumar Hati, Pankaj Kumar Sa, and Banshidhar Majhi,
Intensity Range Based Background Subtraction for
Effective
Object
Detection
IEEE
SIGNAL
PROCESSING LETTERS, VOL. 20, NO. 8, AUGUST
2013
[12]. Jing-Ming Guo, Senior Member, IEEE, Chih-Hsien
Hsia, Member IEEE, Yun-Fu Liu, Student Member,
IEEE, Min-Hsiung Shih, Cheng-Hsin Chang, and JingYu Wu, Fast Background Subtraction Based on a
Multilayer Codebook Model for Moving Object
Detection IEEE TRANSACTIONS ON CIRCUITS
AND SYSTEMS FOR VIDEO TECHNOLOGY, VOL.
23, NO. 10, OCTOBER 2013
[13]. Bo-Hao Chen and Shih-Chia Huang, An Advanced
Moving Object Detection Algorithm for Automatic
Trafc Monitoring in Real-World Limited Bandwidth
Networks
IEEE
TRANSACTIONS
ON
MULTIMEDIA, VOL. 16, NO. 3, APRIL 2014
[14]. Yao-Tsung Yang ,and Ching-Te Chiu, BOOSTED
MULTI-CLASS OBJECT DETECTION WITH
PARALLEL HARDWARE IMPLEMENTATION FOR
REAL-TIME
APPLICATIONS
2014
IEEE
International Conference on Acoustic, Speech and
Signal Processing (ICASSP) 2014
[15]. Ning Liu, Hefeng Wu, and Liang Lin, Hierarchical
Ensemble of Background Models for PTZ-Based Video
Surveillance,
IEEE
TRANSACTIONS
ON
CYBERNETICS, VOL. 45, NO. 1, JANUARY 2015
[16]. Pin Yi Tsai , Yarsun Hsu , Ching-Te Chiu , Tsai-Te
Chu. Accelerating AdaBoost Algorithm Using GPU
for Multi-Object Recognition 2015 IEEE
[17]. Hakki Can Karaimer, Ibrahim Cinaroglu, Yalin
Bastanlar Combining shape-based and gradient-based
classifiers for vehicle classification 2015 IEEE 18th
International Conference on Intelligent Transportation
Systems
[18]. George Kopsiaftis and Konstantinos Karantzalos
VEHICLE DETECTION AND TRAFFIC DENSITY
MONITORING FROM VERY HIGH RESOLUTION
SATELLITE VIDEO DATA, IEEE 2015
[19]. Yuanyuan Wu, Student Member, IEEE, Xiaohai He,
Member, IEEE, and Truong Q. Nguyen, Fellow, IEEE,
Moving Objects Detection with Freely Moving Camera
via Background Motion Subtraction 2015 IEEE

Volume 5, Issue 2, March April 2016

ISSN 2278-6856

[20]. Polani Devi Vara Prasad, Janakani Devi Vara Prasad,

Malijeddi
Murali
ADVANCED
HARDWARE
ARCHITECTUREFOR
ADABOOSTBASED
REALTIMEOBJECT DETECTION, International
Journal of Advanced Research in Computer
Engineering & Technology Volume 1, Issue 6, August
2012
[21]. Deepak Kumar Panda, Student Member, IEEE, and
Sukadev Meher, Member, IEEE Detection of Moving
Objects Using Fuzzy Color Difference Histogram Based
Background Subtraction IEEE SIGNAL PROCESSING
LETTERS, VOL. 23, NO. 1, JANUARY 2016
[22]. Christos Kyrkou, Member, IEEE, Christos-Savvas
Bouganis, Member, IEEE,Theocharis Theocharides,
Senior Member, IEEE, and Marios M. Polycarpou,
Fellow, IEEE Embedded Hardware-Efcient Real-Time
Classication With Cascade Support Vector Machines,
IEEE TRANSACTIONS ON NEURAL NETWORKS
AND LEARNING SYSTEMS, VOL. 27, NO. 1,
JANUARY 2016
[23]. Deepak Kumar Panda, Student Member, IEEE, and
Sukadev Meher, Member, IEE, Detection of Moving
Objects Using Fuzzy Color Difference Histogram Based
Background Subtraction IEEE SIGNAL PROCESSING
LETTERS, VOL. 23, NO. 1, JANUARY 2016
[24]. Rasiq S.M, S.Krishnakumar Parallel Processors for
High Speed Multiple Objects Recognition, 2014 IEEE
[25]. Ning Liu, Hefeng Wu, and Liang Lin Hierarchical
Ensemble of Background Models for PTZ-Based Video
Surveillance
IEEE
TRANSACTIONS
ON
CYBERNETICS, VOL. 45, NO. 1, JANUARY 2015
[26]. Jun Wu Zhitao Xiao , Video Surveillance Object
Recognition Based on Shape and Color Features, 2010
3rd International Congress on Image and Signal
Processing (CISP2010)
[27]. Ren Meng, Zhang Shengbing, Lei Yi , Zhang Meng
CUDA-based Real-time Face Recognition System 2014
IEEE
[28]. Salvi, An Automated Nighttime Vehicle Counting and
Detection System for Trafc Surveillance , 2014
International Conference on Computational Science and
Computational Intelligence
[29]. Sandesh Patil Kiran Talele Suspicious Movement
Detection and Tracking based on Color Histogram ,
2015 International Conference on Communication,
Information & Computing Technology (ICCICT), Jan.
16-17, Mumbai, India
[30]. Zhilu Chen and Xinming Huang, Zhen Ni and Haibo
He, A GPU-Based Real-Time Trafc Sign Detection
and Recognition System , IEEE 2014
[31]. Himani S. Parekh, Darshak G. Thakore , Udesang K.
Jaliya A Survey on Object Detection and Tracking
Methods, International Journal of Innovative Research
in Computer and Communication Engineering Vol. 2,
Issue 2, February 2014

Page 92

International Journal of Emerging Trends & Technology in Computer Science (IJETTCS)

Web Site: www.ijettcs.org Email: [email protected]
Volume 5, Issue 2, March - April 2016

ISSN 2278-6856

[32]. Kirubaraj Ragland P. Tharcis, A Survey on Object

Detection, Classification and Tracking Methods,
International Journal of Engineering Research &
Technology (IJERT) IJERT ISSN: 2278-0181
IJERTV3IS110458 www.ijert.org (This work is licensed
under a Creative Commons Attribution 4.0
International License.) Vol. 3 Issue 11, November-2014
[33]. Pontil, M. ; Center for Biol. & Comput. Learning,
MIT, Cambridge, MA, USA ; Verri, A. Support vector
machines for 3D object recognition IEEE Transactions
on Pattern Analysis and Machine Intelligence Jun 1998
[34]. Dongkyun Kim et al, An FPGA based parallel
Hardware architecture for Real time eye detection,
Journal of Semiconductor Technology and Science,
June 2013
[35]. Shotton, J. ; Toshiba Corp. R&D Center, Kawaski
; Blake, A. ; Cipolla, R. Multiscale Categorical Object
Recognition Using Contour Fragments, Pattern
Analysis and Machine Intelligence, IEEE Transactions
on (Volume:30 , Issue: 7 )
[36]. Alvarez, J.M.A. ; Dept. of Comput. Sci., Univ.
Autonoma de Barcelona, Cerdanyola del Valles, Spain
; opez, A.M. Road Detection Based on Illuminant
Invariance Intelligent Transportation Systems, IEEE
Transactions on (Volume:12 , Issue: 1 )
[37]. Vinaykumar, M. ; Dept. of E.C.E., Nat. Inst. Of
Technol., Warangal, India ; Jatoth, R.K. Performance
evaluation of Alpha-Beta and Kalman filter for object
tracking 2014 International Conference on Advanced
Communication Control and Computing Technologies
(ICACCCT),
[38]. Alper Yilmaz et al, Object Tracking: A Survey, ACM
Computing Surveys, Vol. 38, No. 4, Article 13,
Publication date: December 2006
[39]. Hiroyuki Ukida , High Speed and Wide Range Object
Tracking System Using Pan-Tilt Cameras and Arm
Robot IMTC 2005 - Instrumentation and Measurement
Technology Conference Ottawa, Canada, May 1719,
2005
[40]. Patricia P. Wang, Wei Zhang, Jianguo Li, Yimin
Zhang
REALTIME DETECTION OF SALIENT
MOVING OBJECT: A MULTI-CORE SOLUTION
2008 IEEE
[41]. Dr. Arpita Gopal Sonali Patil Amresh Nikam, A
parallel algorithm for outline capturing system for
object recognition , 2010 IEEE

Volume 5, Issue 2, March April 2016

Page 93