Council For Innovative Research: Efficient Motion Detection Algorithm in Video Sequences
Council For Innovative Research: Efficient Motion Detection Algorithm in Video Sequences
Council For Innovative Research: Efficient Motion Detection Algorithm in Video Sequences
aree.ali@univsul.net
ABSTRACT
Human motion analysis concerns the detection, tracking and recognition of people behaviors, from image
sequences involving humans. A reference frame is initially used and considered as background information. While
a new object enters into the frame, the foreground information and background information are identified using the
reference frame as background model.In this paper, an efficient algorithm is proposed for objects detection in real time
video sequences. The method aims at tracking an object like (human) in motion using background subtraction technique.
The tracked objects are subject to different pre and post image processing in order to extract the most important features
(frame preprocessing is used for shadow removal using morphological operations). They can be used later for recognition
in different security system such as (human detection for surveillance). Experimental results show that the proposed
method is very efficient in terms of reliability and accuracy of detection.
Indexing terms/Keywords
Motion Detection; Object Tracking; Background Subtraction Method; Image Preprocessing, Shadow Removal.
ISSN 2277-3061
I INTRODUCTION
In video sequences, moving objects identification is a critical task in human detection and tracking, video surveillance,
traffic monitoring and analysis. A common approach to identifying the moving objects is background subtraction, where
each video frame is compared against a reference or background model. Pixels in the current frame that deviate
significantly from the background are considered to be moving objects. These foreground pixels are further processed for
object localization and tracking. Since background subtraction is often the first step in many computer vision applications,
it is important that the extracted foreground pixels accurately correspond to the moving objects of interest. Even though
many background subtraction algorithms have been proposed in the literature, the problem of identifying moving objects in
complex environment is still far from being completely solved. There are several problems that a good background
subtraction algorithm must solve correctly [1].
There are many algorithms for motion detection in a continuosus video stream when a camera is fixed; most of them
based on comparing of the current vide frame with one from the previous frames or with something that is called
background. In background subtraction algorithm, detecting the forground objects are the difference between the current
frame and the static frame as a background [2].
framei
backgroundi Threshold
Where Threshold value is used to control the objects in motion with respect to the static background (improve the
subtraction). A reliable and robust background subtraction algorithm should handle:
1. Sudden or gradual illumination changes,
2. High frequency, repetitive motion in the background (such as tree leaves, flags, waves, .), and
3. Long term scene changes (a car is parked for a month).
Simple approach:
a.
b.
c.
Apply a threshold to the absolute difference to get the forground mask (objects of interest).
II RELATED WORK
Over the past years, several papers have been published addressing the evaluation of BS methods In this section some
related research work is described about the motion detection of the objects in a video sequences.
Wren et al. describes the real-time Pnder system for detecting and tracking humans [3]. The background model uses a
Gaussian distribution in the YUV space at each pixel, and the background model is continually updated.
In [4] the intensity di erence between an input frame and a reference image as a multi-modal probability distribution, and
mode detection is performed by using mean shift computation is treated. The mean shift computation is performed in a
fast way using integral images or summed area tables, which gives the method real-time performance in a manner which
is independent of the size of thewindow used.
Paper [5] presents a method to detect and track a human body in a video. First, background subtraction is performed to
detect the foreground object, which involves temporal dierencing of the consecutive frames. After this step, the
classication of the object is based on two approaches: the rst is a codebook approach, and the second involves tracking
of the object and if the object can be tracked successfully, it is considered to be a human.
This paper [6] proposes an adaptive correspondence estimation approach between person detections in a planar scene
not relying on correspondence features as it is the case in many other RANSAC-based approaches. The result is a planar
4294 | P a g e
ISSN 2277-3061
inter-image homography calculated from estimated point correspondences. The approach is self-configurable, adaptive
and provides robustness over time by exploiting temporal and geometric information.
Nijad et al presents a low cost automatic object tracking algorithm suitable for use in real-time video based systems.
The novelty of the proposed system is that it uses a simplified version of the Kanade-Lucas-Tomasi (KLT)
technique to detect features of both continuous and discontinuous nature. As discontinuous feature selection is
subject to noise, and would result in non-optimal feature based object tracking, the authors propose the use of a Kalman
filter for the purpose of seeking optimal estimates in tracking. The integrated tracking system is capable of handling
shadows and is based on a dynamic background subtraction strategy that minimises errors and quickly adapts to
scene changes [7].
Liu et al [8] presents a new approach to real time human detection and tracking in cluttered and dynamic enviroments by
integration of RGB and depth data. They introduced the notion of Point Ensemble Image, which fully encodes both RGB
and depth information from a virtual plan-view perspective, and we reveal that human detection and tracking in 3D space
can be performed very effectively based on this new representation. The proposed human detector is able to take
advantage of depth data by effectively locate physically plausible candidates as a rst step, and then both depth and color
information is made full use of in a supervised learningmanner at the second stage.
IV PROPOSED METHODS
First of all, a general block diagram of the proposed motion detection of human objects in a real time video is given in
figure
1.
A. Motion Detector
Background subtraction technique is used to find the foreground object from the video and then classify it into
categories like human, animal, vehicle etc., based on motion estimation between consecutive frames. The
classification involves tracking of the object and if the object can be tracked successfully, it is considered to be a
human. The algorithm steps for detecting human body in motion are as follow:
Step1: grap the ith frames (8 bit gray scale image)
Step2: grap (i-3) th frame
Step3: Apply background subtraction algorithm (Subtract I from (i-3))
4295 | P a g e
ISSN 2277-3061
Step4: Convert the gray image into binary using thresholding technique
Step5: In the thresh-holding technique a parameter called the brightness threshold (T) is chosen and applied to the
image f[m,n] as follows:
IF f[m,n]T
ELSE
fb[m,n]=object=1
fb[m,n]=background=0
Step6: Apply some morphological operations (erosion and closing) to fill out the holes
Motion Tracker
After obtaining a human body with a motion detector that describe in section A, the center of gravity (COG) of the
binary image is calculated. .The result of this operation is a set of two integers C(cog_x ,cog_y) which determines the
position of the moving object in the given scene.
The COG is calculated by:
cog_x =cog_y + x
(1)
cog_y = cog_y + y
(2)
Total = Total + 1
(3)
for each pixel where x, y is the current pixel location. The resulting COG is then divided by the Total value:
cog_x = cog_x/Total
(4)
cog_y = cog_y/Total
(5)
4296 | P a g e
ISSN 2277-3061
Figure 3 presents the stages of background subtraction method based on object in motion.
Fig 3: top left frame difference top right binary thresholding bottom left bottom right human tracking
Another test is done using multiple human objects in motion as illustrated in figure 4.
Fig 4: top left current frame top right frame difference bottom left binary bottom right human tracking
4297 | P a g e
ISSN 2277-3061
VI CONCLUSION
Efficient human body (single and multiple) detection in motion is presented. The proposed method is overcome the
problem of shadowing due to the lighting condition. The human motion is tracked based on an improved template
matching algorithm between consecutive frames.The experimental results show that the system is very accurate and
efficient for single object detection. The detection accuracy is reached to 99.7% for different lighting condition and various
distances from a camera. While for the multiple object detection the accuracy is decreased to about 98% due to the
occlusion regions. The system can be very much applicable to areas like surveillance and video conferencing.
ACKNOWLEDGMENTS
Our thanks to the member staff of Information Technology Directorate at the University of Sulaimani who have contributed
towards development of this work. I would like also to thank my colleagues at the computer science department for their
help and support to complete this research work.
REFERENCES
[1] Harsha, V., Heena, C., Kajal, S., Shruti, G., Tina, G. and Dashrath, M. 2013. Understanding various Techniques for
Background Subtraction and Implementation of Shadow Detection. IJCTA, Vol 4 (5), (Oct -2013), 822-827.
[2] Marwa, A. 2010. Modified Background Subtraction Algorithm for Motion Detection in Surveillance Systems.
AMARABAC, Vol 1 (2), 2010, 112-123.
[3] C. R. Wren, A. Azarbayejani, T. Darrell, and A. P. Pentland, 1997. Pnder: Real-Time Tracking of the Human Body.
IEEE Transactions on Pattern Analysis and Machine Intelligence, 19(7), 1997, :780785.
[4] C. Beleznai, B. Fruhstuck, and H. Bischof, 2004. Human Detection in Groups using a Fast Mean Shift Procedure.
International Conference on Image Processing, Vol. 1, 2004, 349352.
[5] Jianpeng Zhou and Jack Hoang, 2005. Real Time Robust Human Detection and Tracking System. IEEE Computer
Society Conference on Computer Vision and Pattern Recognition, 3, 2005, 149 149.
[6] M. Thaler and R. Morzinger, 2010. Automatic Inter-Image Homography Estimation from Person Detections. In
Advanced Video and Signal Based Surveillance (AVSS),2010 Seventh IEEE International Conference on, 2010, 456
461.
[7] Nijad Al-Najdawi, Sara Tedmori, Eran Edirisinghe, and Helmut Bez, 2012. An Automated Real-Time People Tracking
System Based on KLT Features Detection. The International Arab Journal of Information Technology, Vol. 9, No. 1,
January 2012, 100-107.
[8] Jun Liu, Ye Liu, Ying Cui, and Yan Qiu Chen, Real Time Human Detection and Tracking in Complex Enviroments
using Single RGBD Camera. IEEE conference ICIP 2013, 3088-3092.
[9] Spector, A. Z. 1989. Achieving application requirements. In Distributed Systems, S. Mullender
4298 | P a g e