A Survey On Tracking Moving Objects Using Various Algorithms

A Survey On Tracking Moving Objects Using Various Algorithms Page 21
International Journal for Modern Trends in Science and Technology
ISSN: 2455-3778 |Volume No: 02 | Issue No: 02 | February 2016
A Survey On Tracking Moving Objects Using
Various Algorithms
Vaishnavi.S1, Chitti Babu.K2 and Kavitha.C3
1sevavaishnavi@gmail.com, Dept. of CSE, R.M.K. College of Engineering and Technology, Tamilnadu-601206, INDIA.
2mailtocbabu@gmail.com, Dept. of CSE, R.M.K. College of Engineering and Technology, Tamilnadu-601206, INDIA.
3kavitha4cse@gmail.com, Dept. of CSE, R.M.K. College of Engineering and Technology, Tamilnadu-601206, INDIA.
Article History ABSTRACT
Received on: 15-02-2016
Accepted on: 19-02-2016 Sparse representation has been applied to the object tracking
problem. Mining the self- similarities between particles via
multitask learning can improve tracking performance. How-ever,
some particles may be different from others when they are sampled
from a large region. Imposing all particles share the same structure
may degrade the results. To overcome this problem, we propose a
tracking algorithm based on robust multitask sparse representation
(RMTT) in this letter. When we learn the particle representations, we
decompose the sparse coefficient matrix into two parts in our algorithm.
Joint sparse regularization is imposed on one coefficient matrix while
element-wise sparse regularization is imposed on another matrix. The
former regularization exploits self-similarities of particles while the later
one considers the differences between them.
Published on: 25-02-2016
Keyword
Element-wise sparse,
Regularization,
Joint sparse,
Regularization,
Sparse representation.
Copyright © 2015 International Journal for Modern Trends in Science and Technology
All rights reserved.

I. INTRODUCTION
Visual object tracking is an important topic in
computer vision and has many applications
including intelligent surveillance, human computer
interface (HCI), augmented reality (AR), etc.
Although great processes have been made and
many tracking algorithms have been proposed but
it remains an open problem to design a robust
tracker in the real-world scenarios due to severe
occlusions, large appearance changes, illumination
changes, background clutter and abrupt motion.
Recently, sparse representation and compress
sensing have been applied to the object tracking
problem. In the L1 tracker proposed by Mei and
Ling, each candidate is sparsely represented by the
target templates and trial templates. This
representation is robust to illumination changes
and partial occlusion, the stable signal recovery
capability via the L1 norm minimization. To solve
the L1 norm minimization efficiently, we adopt the
accelerated proximal gradient (APG) approach and
they claim the L1APG tracker run in the real-time.
When particles are sampled from a large region and
the brutal enforcement of the same structure on all
particles, e.g. joint-sparsity, low-rank, may degrade
the results. To overcome this problem, we propose
our tracking algorithm based on robust multitask
learning (RMTT). In our algorithm, we decompose
the sparse coefficient matrix into two parts. A joint
sparse regularization is imposed on, corresponding
to the shared structure while an element-wise
sparse regularization is imposed on which
corresponds to the non-shared features. The joint
sparsity exploits the similarities of particles while
the element-wise sparsity considers the differences
between candidates. Many experiments on the
benchmark datasets show the superior
performance over the state-of-art algorithms.
Tracking is an optical method that employs
tracking and image registration techniques for
accurate 2D and 3D measurements of changes in
images. This is often used to measure deformation
(engineering), displacement, strain, and optical
flow, but it is widely applied in many areas of
science and engineering. In its simplest form,
tracking can be defined as the problem of
estimating the trajectory of an object in the image
plane as it moves around a scene. In other words, a
tracker assigns consistent labels to the tracked
objects in different frames of a video. Additionally,
depending on the tracking domain, a tracker can
also provide object-centric information, such as
orientation, area, or shape of an object. One very
common application is for measuring the motion of
an optical mouse.
II. LITERATURE SURVEY
Most tracking-by-detection algorithms train
discriminative classifiers to separate target objects
from their surrounding background. In this setting
[15], noisy samples are likely to be included when
they are not properly sampled, thereby causing
visual drift. The multiple instance learning (MIL)
paradigm has been recently applied to alleviate this
problem. However, important prior information of
instance labels and the most correct positive
instance (i.e., the tracking result in the current
frame) can be exploited using a novel formulation
much simpler than an MIL approach. In this paper,
we show that integrating such prior information
into a supervised learning algorithm can handle
visual drift more effectively and efficiently than the
existing MIL tracker.
Sparse representation scheme is very
influential in visual tracking field [8]. These L1
trackers obtain robustness by finding the target
with the minimum reconstruction error via L1
norm minimization problem. However, the high
computational burden of L1 minimization and
absence of effective model for appearance changes
may hamper its application in real world sceneries.
In this research, we present a fast and robust
tracking method that exploits a fast memory
gradient pursuit algorithm (FMGP) with sparse
representation scheme in a Bayesian framework to
accelerate the L1 minimization process.
Combining multiple observation views
has proven beneficial for tracking. In paper [17], we
cast tracking as a novel multi-task multi-view
sparse learning problem and exploit the cues from
multiple views including various types of visual
features, such as intensity, color, and edge, where
each feature observation can be sparsely
represented by a linear combination of atoms from
an adaptive feature dictionary. The proposed
method is integrated in a particle filter framework
where every view in each particle is regarded as an
individual task. We jointly consider the underlying

relationship between tasks across different views
and different particles, and tackle it in a unified
robust multi-task formulation.
The use of multiple features for tracking has
been proved as an effective approach because
limitation of each feature could be compensated
[10]. Since different types of variations such as
illumination, occlusion and pose may happen in a
video sequence, especially long sequence videos,
how to dynamically select the appropriate features
is one of the key problems in this approach. To
address this issue in multicue visual tracking, this
paper proposes a new joint sparse representation
model for robust feature-level fusion. The proposed
method dynamically removes unreliable features to
be fused for tracking by using the advantages of
sparse representation. As a result, robust tracking
performance is obtained. Experimental results on
publicly available videos show that the proposed
method outperforms both existing sparse
representation based and fusion-based trackers.
In paper [12], we formulate object tracking in
a particle filter framework as a multi-task sparse
learning problem, which we denote as Multitask
Tracking (MTT). Since we model particles as linear
combinations of dictionary templates that are
updated dynamically, learning the representation
of each particle is considered a single task in
MTT.By employing popular sparsity-inducing `p;q
mixed norms, we regularize the representation
problem to enforce joint sparsity and learn the
particle representations together. As compared to
previous methods that handle particles
independently, our results demonstrate that
mining the interdependencies between particles
improves tracking performance and overall
computational complexity.
We address the problem of visual
classification with multiple features and/or
multiple instances. Motivated by the recent success
of multitask joint covariate selection [3], we
formulate this problem as a multitask joint sparse
representation model to combine the strength of
multiple features and/or instances for recognition.
A joint sparsity-inducing norm is utilized to enforce
class-level joint sparsity patterns among the
multiple representation vectors. The proposed
model can be efficiently optimized by a proximal
gradient method. Furthermore, we extend our
method to the setup where features are described
in kernel matrices. We then investigate into two
applications of our method to visual classification:
1) fusing multiple kernel features for object
categorization and 2) robust face recognition in
video with an ensemble of query images.
We propose a new particle-filter based
tracking algorithm [14] that exploits the
relationship between particles (candidate targets).
By representing particles as sparse linear
combinations of dictionary templates, this
algorithm capitalizes on the inherent low-rank
structure of particle representations that are
learned jointly. This low-rank sparse tracker (LRST)
has a number of attractive properties. (1) Since
LRST adaptively updates dictionary templates, it
can handle significant changes in appearance due
to variations in illumination, pose, scale, etc. (2)
The linear representation in LRST explicitly
incorporates background templates in the
dictionary and a sparse error term, which enables
LRST to address the tracking drift problem and to
be robust against occlusion respectively. (3) LRST
is computationally attractive, since the low-rank
learning problem can be efficiently solved as a
sequence of closed form update operations.
Object tracking in a particle filter
framework as a structured multi-task sparse
learning problem [13], which we denote as
Structured Multi-Task Tracking (S-MTT). Since
we model particles as linear combinations of
dictionary templates that are updated
dynamically, learning the representation of each
particle is considered a single task in Multi-Task
Tracking (MTT). By employing popular sparsity-
inducing _p,q mixed norms (specifically p ∈{2,∞}
and q = 1), we regularize the representation
problem to enforce joint sparsity and learn the
particle representations together. As compared
to previous methods that handle particles
independently, our results demonstrate that
mining the interdependencies between particles
improves tracking performance and overall
computational complexity.
A. Disadvantages
Joint-Sparsity, Low-Rank, may degrade the
results. Both MTT and LRST show the superior

performance over both L1 and L1APG trackers. The
background is cluttered and complex and the
target were occluded. When a similar region
appeared, L1apg was confused and hijacked by this
region. When the occlusion was happened, MTT
introduced large error and almost lost the target
location.
III. PROPOSED SYSTEM
Sparse representation has been applied to
the object tracking problem. Mining the self-
similarities between particles via multitask learning
can improve tracking performance. However, some
particles may be different from others when they
are sampled from a large region. Imposing all
particles share the same structure may degrade the
results. To overcome this problem, we propose a
tracking algorithm based on robust multitask
sparse representation (RMTT). This is the main
problem that they do not consider difference
between particles when exploiting the
dependencies. When we learn the particle
representations, we decompose the sparse
coefficient matrix into two parts in our algorithm.
Joint sparse regularization is imposed on one
coefficient matrix while element-wise sparse
regularization is imposed on another matrix. The
former regularization exploits self-similarities of
particles while the later one considers the
differences between them. The challenges of these
videos include illumination variation , partial
occlusion , pose change, background clutter
and scale variation . Experiments on the
benchmark data show the superior performance
over other state-of-art algorithms.
Figure 1: Object tracking in video
A. Advantages
This representation is robust to illumination
changes and partial occlusion. This algorithm
captures a common feature among relevant tasks
and identifies outlier tasks. In real world scenarios
it has many applications in intelligent surveillance,
human computer interface.
Figure 2: Tracking object
IV. CONCLUSION
In this paper, we have proposed an
object tracking algorithm via robust multitask
sparse representation. When we learn the sparse
coefficient matrix, we decompose it into two
halves. Joint sparsity regularization is imposed
on one coefficient matrix and element-wise
sparsity regularization is imposed on the other.
This can exploits the dependencies between
patches while considering the differences
between them. Experiments on challenging
video sequences show that our tracking
algorithm performs better than several state-of-
the-art algorithms.
V. FUTURE ENHANCEMENT
We will investigate that the tracking of
the object by an adaptation of the Hough
Transform, which is used in the feature
extraction procedure to interpret scanned
segments as primitive features, defined by
geometric evidence (points, lines, circles and
blobs), and high-level features, generally
referred to as landmarks (corners, columns,
doors, etc.). The classification system uses
features data, some heuristic rules, and data
from a filter based tracking system to classify
multiple objects.
ACKNOWLEDGMENT
The authors would like to thank the anonymous
reviewers for their constructive and useful
comments.
REFERENCES
[1] S. Baker and I. Matthews, “Lucas-Kanade 20 years
on: A unifying framework,” Int. J. Comput. Vis., vol. 56,
no. 3, pp. 221–255, 2004.
[2] A. Adam, E. Rivlin, and I. Shimshoni, “Robust
fragments-based tracking using the integral histogram,”
in CVPR, 2006.

[3] Xiao-Tong Yuan, Xiaobai Liu, and Shuicheng
Yan"Visual Classification With Multitask Joint Sparse
Representation",2012.
[4] S. Hare, A. Saffari, and P. H. Torr, “Struck:
Structured output tracking with kernels,” in ICCV, 2011.
[5] J. Kwon and K. M. Lee, “Visual tracking
decomposition,” in CVPR, 2010.
[6] Y. Bai and M. Tang, “Robust tracking via weakly
supervised ranking svm,” in CVPR, 2012.
[7] M. Tang and X. Peng, “Robust tracking with
discriminative ranking lists,” IEEE Trans. Image
Process., vol. 21, no. 7, pp. 3273–3281, Jul. 2012.
[8]Qiang Guo and Chengdong Wu1,"Fast Visual Tracking
using memory gradient pursuit algorithm",2014.
[9] X. Mei and H. Ling, “Robust visual tracking using l1
minimization,” in ICCV, 2009.
[10] Xiangyuan,Lan Andy J Ma,Pong C Yuen,"Multi-Cue
Visual Tracking Using Robust Feature-Level Fusion
Based on Joint Sparse Representation",2014
[11] P. Tseng,“On accelerated proximal gradient methods
for convex-concave optimization,” SIAM J. Optim., 2008.
[12] T. Zhang, B. Ghanem, S. Liu, and N. Ahuja, “Robust
visual tracking via multi-task sparse learning,” in CVPR,
2012.
[13] T. Zhang, B. Ghanem, S. Liu, and N. Ahuja, “Robust
visual tracking via structuredmulti-task sparse learning,”
Int. J. Comput. Vis., vol. 101, pp. 367–383, 2013.
[14] T. Zhang, B. Ghanem, S. Liu, and N. Ahuja, “Low-
rank sparse learning for robust visual tracking,” in
ECCV, 2012, pp. 470–484.
[15] Kaihua Zhang, Lei Zhang, IEEE, and Ming-Hsuan
Yang"Real-Time Object Tracking via Online
Discriminative Feature Selection",2013.
[16] P. Gong, J. Ye, and C. Zhang, “Robust multi-task
feature learning,” in SIGKDD, 2012.
[17] Zhibin Hong1, Xue Mei2, Danil Prokhorov2, and
Dacheng Tao1,"Tracking via Robust Multi-Task Multi-
View Joint Sparse Representation",2014.

A Survey On Tracking Moving Objects Using Various Algorithms

More Related Content

What's hot (20)

Viewers also liked (13)

Similar to A Survey On Tracking Moving Objects Using Various Algorithms (20)

Recently uploaded (20)

A Survey On Tracking Moving Objects Using Various Algorithms