F. Erden and A. E.

Çetin: Hand Gesture Based Remote Control System Using Infrared Sensors and a Camera 675

Hand Gesture Based Remote Control System

Using Infrared Sensors and a Camera
Fatih Erden and A. Enis Çetin, Fellow, IEEE

Abstract — In this paper, a multimodal hand gesture Video based hand gesture recognition is an active research
detection and recognition system using differential Pyro- area [7], [8]. Although these approaches provide flexibility in
electric Infrared (PIR) sensors and a regular camera is terms of recognition distance and accuracy, they may make
described. Any movement within the viewing range of the false detections due to the illumination changes and reflection.
differential PIR sensors are first detected by the sensors and As pointed out by Wachs et al. [9], there are only a few
then checked if it is due to a hand gesture or not by video practical vision based hand gesture recognition systems.
analysis. If the movement is due to a hand, one-dimensional Current Pyro-electric Infrared (PIR) hand gesture
continuous-time signals extracted from the PIR sensors are recognition systems are all based on the on/off decisions of
used to classify/recognize the hand movements in real-time. the analog circuitry of the PIR sensor [10]-[12]. Wojtczuk et
Classification of different hand gestures by using the al. [13] use a 4x4 PIR sensor array to obtain a switching
differential PIR sensors is carried out by a new winner-take- pattern of PIR sensors to increase the recognition rates. The
all (WTA) hash based recognition method. Jaccard distance is
4x4 PIR sensor array has the capability to recognize a number
used to compare the WTA hash codes extracted from 1-D
of hand gestures but it cannot distinguish face and body
differential infrared sensor signals. It is experimentally shown
that the multimodal system achieves higher recognition rates gestures from hand gestures, because PIR sensors respond to
than the system based on only the on/off decisions of the all hot bodies in their viewing range. Besides an additional
analog circuitry of the PIR sensors1. camera used for hand detection, the novelty of the multimodal
system presented here is the use of continuous-time real-
Index Terms — Hand gesture recognition, infrared sensors, valued signals that PIR sensors produce during an action. As a
sensor fusion, wavelet transform, winner-take-all (WTA) hash result it achieves higher recognition rates than the 4x4 PIR
method. sensor array [13] due to the novel use of the analog PIR
sensor signals.
I. INTRODUCTION The proposed multimodal system consists of a differential
Studies on hand and face gesture recognition started in the PIR sensor array and an ordinary camera for hand gesture
early 1990s and continues ever since [1]. These studies have recognition. Data obtained from the PIR sensor array and the
many potential applications in the areas of human-computer camera are transferred to a computer and then processed
interaction (HCI), virtual reality (VR), remote control (RC) together in real-time. Once any kind of motion is detected by
and industry [2]. Remote and contact-free control of electrical one of the PIR sensors, the motion is checked whether it is
appliances has become a desirable feature in consumer and due to a hand or not by the camera. If the source of the motion
industrial electronics. Contact-free user interfaces are also is a hand gesture, data received from the three different PIR
useful to reduce the hygiene risks in public places. sensors are evaluated at the same time to classify the hand
This paper focuses on hand gesture recognition using gestures e.g., right-to-left/left-to-right, upward/downward
infrared (IR) sensors. Hand gesture recognition systems may hand motions etc. Classification of the hand gestures by the
be classified into three groups: systems using hand-held PIR sensor array is carried out by a new winner-take-all
pointing devices, systems using wearable sensors and systems (WTA) hash based method. This multimodal solution to the
doing two dimensional (2-D) image analysis [3]. Generally, it hand gesture detection and recognition problem is a good
may not be practical to utilize systems requiring gloves or alternative to the existing methods because of its accuracy,
similar equipment for gesture recognition [4]-[6]. It is low cost and low power consumption.
desirable to have natural hand-gestures to control appliances This paper extends an earlier study [14] which describes a
and TV sets. hand gesture based remote control system using two infrared
sensors and camera. The earlier study presents the results
1
produced by the camera-only and the multi-modal systems and
Fatih Erden is with the Department of Electrical and Electronics
Engineering, Hacettepe University (Beytepe Campus), Ankara 06800 reports the improvements made to recognition accuracy of the
TURKEY (e-mail: erden@[Link]). hand gestures by the multi-modal system. In this paper, the
A. Enis Çetin is with the Department of Electrical and Electronics camera is just used for the detection of a hand and the hand
Engineering, Bilkent University, Ankara 06800 TURKEY (e-mail:
enis@[Link]).
gestures are classified using only the sensor array which in
this case consists of three differential PIR sensors. By
Contributed Paper
Manuscript received 09/02/14
Current version published 01/09/15
Electronic version published 01/09/15. 0098 3063/14/$20.00 © 2014 IEEE
Authorized licensed use limited to: GMR Institute of Technology. Downloaded on January 24,2025 at 08:50:57 UTC from IEEE Xplore. Restrictions apply.
676 IEEE Transactions on Consumer Electronics, Vol. 60, No. 4, November 2014

Fig. 1. The circuit diagram for capturing an analog output signal from a PIR sensor.

inserting one more PIR sensor to the system and modifying Jaccard distance between these binary codes and the reference
the WTA hash based classification algorithm accordingly, codes created during training. The action is classified
the new system is now capable of recognizing an additional according to the training data and using the Jaccard distance
set of hand movements, i.e. besides vertical and horizontal metric.
hand motions it can recognize the circular
(clockwise/counter-clockwise) motions of a hand as well. A. Operating Principles of the PIR Sensor System and Data
Furthermore, with the new setup and updates in the methods Acquisition
used in both video and PIR sensors signals analysis the A differential PIR sensor basically measures the difference
proposed system achieves higher recognition accuracy for a of infrared radiation density between the two pyro-electric
greater set of hand motions. elements inside. The elements connected in parallel cancel the
Operating principles of a differential PIR sensor and the normal temperature alterations and the changes caused by
PIR sensor array based hand gesture recognition method are airflow. When these elements are exposed to the same amount
described in Section II. WTA hash coding based sensor of infrared radiation they cancel each other and the sensor
decision fusion is described in Section II-B. Video analysis produces a zero-output. Thus the analog circuitry of the PIR
carried out for hand detection is presented in Section III. sensor can reject false detections very effectively.
Experiments and results are presented in Section IV. Commercially available PIR motion detector circuits
produce binary outputs. However, it is possible to capture a
II. HAND GESTURE DETECTION AND RECOGNITION USING continuous-time analog signal representing the amplitude of
DIFFERENTIAL PIR SENSOR ARRAY the voltage signal which is the transient behavior of the
Differential PIR sensors give response to the change of circuit. The corresponding circuit for capturing an analog
infrared radiation in their viewing range. But this change may output signal from the PIR sensor is shown in Fig. 1. The
be caused by the motion of a hand, head or the whole body. circuit consists of four operational amplifiers (op amps), U1A,
To solve this problem, the proposed system first determines U1B, U1C and U1D. U1A and U1B constitute a two stage
for the presence of any motion in the interested area with the amplifier circuit whereas U1C and U1D couple behaves as a
help of the PIR sensors. Whenever the analog decision comparator. The very-low amplitude raw output at the 2nd pin
circuitry of one of the PIR sensors determines motion and of the PIR sensor is amplified through the two stage amplifier
video analysis decides that the motion is due to a hand circuit. The amplified signal at the output of U1B is fed into
gesture, the PIR sensor signals are recorded and analyzed in the comparator structure which outputs a binary signal, either
real-time. Hand gestures classified by the differential PIR 0 V or 5 V. Instead of using binary output in the original
sensor array-only system are right-to-left, left-to-right, version of the PIR sensor read-out circuit, the analog output
upward, downward, clockwise and counter-clockwise signal at the output of the 2nd op amp U1B is captured directly.
motions. The analog output signal is digitized using a microcontroller
Feature parameters extracted from the PIR sensor signals with a sampling rate of 100 Hz and transferred to a general-
are based on wavelet transform. Wavelet domain feature purpose computer for further processing. A typical sampled
vectors are transformed into binary codes using WTA hash differential PIR sensor output signal for no activity case using
algorithm. A similarity metric is proposed by calculating the 8 bit quantization is shown in Fig. 2.

Authorized licensed use limited to: GMR Institute of Technology. Downloaded on January 24,2025 at 08:50:57 UTC from IEEE Xplore. Restrictions apply.
F. Erden and A. E. Çetin: Hand Gesture Based Remote Control System Using Infrared Sensors and a Camera 677

B. Processing Sensor Output Signals and Decision corresponding to n -th data windows of the PIR sensor signals
Mechanism in the right side, left side and the one on top, respectively.
In this approach, wavelet based signal processing methods Vectors wr ,n , wl ,n and wu ,n are concatenated and a vector ws ,n , of
are used to extract features from sensor signals. Wavelet domain length 300 for the n -th window is formed. Afterwards this
analysis provides robustness to variations in the sensor signal feature vector, which represents all members of the PIR sensor
caused by temperature changes in the environment. In order to array at the same time, is transformed into binary codes using
keep the computational cost of the detection low, Lagrange the WTA hash method. WTA hashing provides a way to
filters are used to compute wavelet coefficients. convert arbitrary feature vectors into compact binary codes.
These codes are resilient to small perturbations in the feature
vector. They preserve the rank correlation and can be easily
calculated [16]. Computing of a WTA code C X , corresponding
to a feature vector w s , n is explained in Algorithm I.

ALGORITHM I
WTA HASH ALGORITHM
a. Generate random permutation matrices- i s, each
of size M  M , i  1, 2,3,..., h .
( M : the length of the vector ws ,n )
b. For i  0: h  1
* Wi   i ws ,n
* select first K items of Wi
* find the index of the maximum item
* convert to binary code- c xi
Fig. 2. A typical PIR sensor output signal when there is no acitivity within c. C X  [cx , cx ,..., cx ]
0 1 h 1
its viewing range (sampled at 100 Hz)

Let x[n] be a sampled version of the signal produced by Random permutation matrix- i has exactly one entry 1 in
one of the PIR sensors. Wavelet coefficients are obtained after each row and each column and 0s elsewhere. i s are
a single stage sub-band decomposition corresponds to [25 Hz, generated once at the beginning of the algorithm and then
50 Hz] frequency band information of the original sensor used to permute the wavelet coefficient sequences
output signal x[n] because the sampling rate is 100 Hz. In this corresponding to each data window.
single stage sub-band decomposition, output signal is filtered
with an integer arithmetic high-pass filter corresponding to
Lagrange wavelets [15] followed by decimation by 2. The
transfer function of the high-pass filter- H ( z) is given by:
1 1 1 1
H ( z)   ( z  z ). (1)
2 4
The wavelet signals obtained from the two members of the
PIR sensor array due to a right-to-left hand motion are shown
in Fig. 3. In the hand gesture recognition system, viewing
range of the PIR sensors is directionally arranged so that left-
to-right, right-to-left, upward, downward, clockwise and
counter-clockwise hand movements are determined more
accurately. As shown in Fig. 3, the PIR on the right responds
earlier than the PIR on the left to a hand motion from right to
left. The hand enters the viewing range of the right-most PIR
first due to the directional arrangement of the PIR sensors. Fig. 3. Wavelet transform of the two PIR sensors signals due to a right-to-
Sampled signals from each differential PIR sensor are divided left hand motion.

into time windows of length 200 samples covering a 2 seconds In the classification process of 1-D signals received from the
period and wavelet coefficient sequences of length 100 PIR sensors, the method described by Dean et al. [17], which
corresponding to each window are computed. The demonstrator calculates a WTA code based Hamming distance metric to find the
PIR sensor array system consists of three PIR sensors each of similarity of images, is used. Hamming distance value depends on
which is one corner of a triangle. Let wr ,n [k ] , wl ,n [k ] and wu ,n [k ] the length of the code sequence. Instead, Jaccard distance metric
k  (1,...,100) represent the wavelet coefficient sequences which takes values in the interval [0, 1] independent of the length

Authorized licensed use limited to: GMR Institute of Technology. Downloaded on January 24,2025 at 08:50:57 UTC from IEEE Xplore. Restrictions apply.
678 IEEE Transactions on Consumer Electronics, Vol. 60, No. 4, November 2014

of the code sequence is used. The class affiliation of each data mask. In dilation phase bright areas within the image are
window is determined by computing the Jaccard distance between grown by scanning a circle shaped kernel over the image.
the binary code of the data window and the representative member During the erosion phase the same circle shaped kernel is
of each action class. While calculating the Jaccard distance used. As a result bright areas of the image become thinner
between any C1 and C 2 binary code sequences of the same length, after the erosion procedure making the fingers more visible.
the contingency table is formed first (Table I). B. Convex Hull-Defect Analysis
Parameters p , q, r and s in Table I represent the number of
In this stage, contours extracted from the binary hand mask
cases C1,i and C 2,i are both 1, C1,i is 0 and C 2,i is 1, C1,i is 1 and are processed using the convex hull-defect analysis. After
C 2,i is 0 and C1,i and C 2,i are both 0, respectively. Depending calculating the area of the extracted contours, the largest area
upon these values, the Jaccard distance between C1 and C 2 is found and assumed to be a ‘hand’. By lifting his or her hand
closer to the camera than his or her face, the user clearly
sequences is calculated as follows:
makes his or her intention to control the system as illustrated
qr in Fig. 4.
d C1 ,C2  (2)
pqr

TABLE I
CONTINGENCY TABLE
C1
1 0
1 p q
C2
0 r s

Jaccard distances between the WTA code of ws ,n and the

codes determined during training phase for each model are
calculated and the model yielding the smallest distance is
reported as the result of the analysis for the n -th data window.

III. HAND DETECTION USING VIDEO ANALYSIS

Motion detected by the PIR sensors may be due to the
movement of a head or whole body of the viewer. Therefore,
identification of hand is critical to reduce false alarms.
Whenever the analog decision circuitry of the PIR sensor Fig. 4. Hand image and the corresponding convex hull-defect and contour
lines.
detects motion in the viewing range of one of the PIR sensors,
video analysis starts. Standard video analysis methods are used A pre-defined hand gesture further reduces false alarms.
to detect hand within the range of the regular camera. Since the For example the user may open his or her fingers as shown in
video analysis is controlled by PIR sensors, the resulting system Fig. 4. This is an unnatural hand gesture therefore it can be
is a low-power consuming system. used as a remote control signal to the appliance. If the system
First, the presence of a raised hand in the viewing range of recognizes more than three fingers, it starts the PIR signal
the camera is detected. If it is a hand it is possible to recognize analysis.
the following hand gestures: (i) right-to-left, (ii) left-to-right, Convex defect analysis algorithm proposed by Graham and
(iii) upward, (iv) downward, (v) clockwise, (vi) counter- Yao [19] provides three important parameters: start point, end
clockwise hand motions, (vii) open, (viii) closed hand decisions point and depth information. These parameters represent
and (ix) counting the number of open fingers. This is useful to respectively, the point where defect starts, where it ends and
change the channels in TV sets and/or set top boxes. One can go the longest distance between these start and end points. The
to the next channel or the previous one and raise or lower the points and the lines are shown on a hand image in Fig 4. The
sound volume by waving his or her hand. number of depth lines above a pre-defined length gives the
A. Skin Detection Using Classical Color Spaces number of open fingers. If the number of open fingers counted
Skin colored regions are first detected and then convex is zero, the hand is closed, and if it is one or greater than one,
hull-defect analysis is performed to recognize the above hand the hand is considered open.
gestures. HSV and YCbCr color spaces are, in general, used
for skin detection [18]. In this article YCbCr color space is IV. EXPERIMENTS AND RESULTS
used, because it is computationally efficient to obtain the The multimodal demonstrator system consisting of three
YCbCr components from the RGB video color data. differential PIR sensors and a camera is shown in Fig. 5. It can
After having the appropriate skin mask, morphological detect hand and recognize left-to-right, right-to-left, clockwise
operations, dilation and erosion, are applied to have a cleaner and counter-clockwise hand gestures. It is also possible to

Authorized licensed use limited to: GMR Institute of Technology. Downloaded on January 24,2025 at 08:50:57 UTC from IEEE Xplore. Restrictions apply.
F. Erden and A. E. Çetin: Hand Gesture Based Remote Control System Using Infrared Sensors and a Camera 679

detect upward/downward motions as well by adding one more multimodal system is able to recognize circular motions, i.e.
PIR sensor aligned with the top PIR sensor into the system. clockwise and counter-clockwise hand motions, as well with a
As pointed in Section III, whenever the user wants to success rate of 94.3%.
interact with an electrical appliance, he or she raises his or her
hand with open fingers in front of the camera. The user can V. CONCLUSION
end the controlling action when he or she makes a fist. After In this article a multimodal hand gesture detection and
detection of a fist by the camera, the multimodal system goes recognition system is presented. Since both infrared and
to standby mode and if the user wants to give more visible range information is used, the proposed system is more
commands, he activates the system again by showing his or accurate than IR-only and less power consuming than camera-
her hands with open fingers in front of the camera. only systems.
A novel WTA code based sensor fusion algorithm is also
presented for 1-D PIR sensor signal processing. The algorithm
fuses the data coming from the different PIR sensors in an
automatic manner to determine left-to-right, right-to-left,
upward, downward, clockwise and counter-clockwise motions.
A Jaccard distance based metric is used to classify the hash
codes of feature vectors extracted from sensor signals.

REFERENCES
[1] J. M. Rehg and T. Kanade, “Digiteyes: Vision-based hand tracking for
human-computer interaction,” in Proc. IEEE Workshop on Motion of
Non-Rigid and Articulated Objects, Texas, USA, pp. 16-22, Nov. 1994.
[2] W. T. Freeman and C. D. Weissman, “Hand gesture machine control
system,” Computer Integrated Manufacturing Systems, vol. 10, no. 2,
pp.175-175, May 1997.
[3] M. Ishikawa and H. Matsumura, “Recognition of a hand-gesture based
on self-organization using a DataGlove,” in Proc. of the 6th
International Conference on Neural Information Processing, vol. 2, pp.
739-745, Perth, Australia, Nov. 1999.
Fig. 5. Setup of the hand gesture detection and recognition system (3 [4] Y. Han, “A low-cost visual motion data glove as an input device to
differential PIR sensors and a laptop’s camera) interpret human hand gestures,” IEEE Trans. Consumer Electron., vol.
56, no.2, pp. 501-509, May 2010.
The system can detect and recognize hand gestures up to [5] O. Z. Ozer, O. Ozun, C. O. Tuzel, V. Atalay, and A. E. Cetin, “Vision-
1.5 meters. It is assumed that the distance between the user based single-stroke character recognition for wearable computing,”
and the TV set is about 1.5 to 2 meters. It is also possible to IEEE Intell. Syst., vol. 16, no. 3, pp. 33-37, Jan./Feb. 2001.
[6] K. Gill, S. H. Yang, F. Yao, and X. Lu, “A zigbee-based home
increase the range using PIR sensors with more directional automation system,” IEEE Trans. Consumer Electron., vol. 55, no.2, pp.
selectivity and range. There is a Fresnel lens in front of most 422-430, May 2009.
PIR sensors. The quality of the lens improves the range of the [7] A. Erdem, E. Erdem, Y. Yardimci, V. Atalay, and A. E. Çetin,
“Computer vision based mouse,” in Proc. IEEE International
sensor. Conference on Acoustics, Speech, and Signal Processing, Florida, USA,
vol. 4, pp. IV-4178, May 2002.
TABLE II [8] D. Lee and Y. Park, “Vision-based remote control system by motion
CLASSIFICATION RESULTS FOR 312 LEFT-TO-RIGHT/RIGHT-TO-LEFT, 298 detection and open finger counting,” IEEE Trans. Consumer Electron.,
UPWARD/DOWNWARD AND 282 CLOCKWISE/COUNTER-CLOCKWISE HAND vol. 55, no.4, pp. 2308-2313, Nov. 2009.
MOTIONS BELONGING TO SIX DIFFERENT USERS. [9] J. P. Wachs, M. Kölsch, H. Stern, and Y. Edan, “Vision-based hand-
gesture applications,” Communications of the ACM, vol. 54, no. 2, pp.
Number of False
Direction Accuracy (%) 60-71, Jan. 2011.
tests detections
[10] R. Samson and T. R. Winings, “Automated dispenser for disinfectant
left/right 312 4 98.7 with proximity sensor,” U.S. Patent 5,695,091, Dec. 9, 1997.
up/down 298 7 97.6 [11] F. Erden et al., “Wavelet based flickering flame detector using differential
circular 282 16 94.3 PIR sensors,” Fire Safety Journal, vol. 53, pp. 13-18, Oct. 2012.
[12] Y. W. Bai, Z. L. Xie, and Z. H. Li, “Design and implementation of a
Classification results for 312 left-to-right/right-to-left, 298 home embedded surveillance system with ultra-low alert power,” IEEE
Trans. Consumer Electron., vol. 57, no.1, pp. 153-159, Feb. 2011.
upward/downward and 282 clockwise/counter-clockwise hand [13] P. Wojtczuk, A. Armitage, T. D. Binnie, and T. Chamberlain,
motions belonging to six different users sitting in front of the “Recognition of simple gestures using a PIR sensor array,” Sensors &
multimodal system at a distance of 2 meters are summarized Transducers, vol. 14, no. 1, pp. 83-94, Mar. 2012.
[14] F. Erden, A. S. Bingöl, and A. E. Çetin, “Hand gesture recognition using
in Table II. The results in the second row of Table II are two differential PIR sensors and a camera,” in Proc. IEEE 22nd Signal
obtained by rotating the PIR sensor array system by 90 Processing and Communications Applications Conference, Trabzon,
degrees. The accuracy reported by Wojtczuk et al. [13] for Turkey, pp. 349-352, Apr. 2014.
[15] C. W. Kim, R. Ansari, and A. E. Cetin, “A class of linear-phase regular
upward/downward, left-to-right/right-to-left cases is 92.6%. biorthogonal wavelets,” in Proc. IEEE International Conference on
On the other hand the overall accuracy of the proposed Acoustics, Speech, and Signal Processing, San Francisco, USA, vol. 4,
multimodal system for these hand motions is 98.2%. The pp. 673-676, Mar. 1992.

Authorized licensed use limited to: GMR Institute of Technology. Downloaded on January 24,2025 at 08:50:57 UTC from IEEE Xplore. Restrictions apply.
680 IEEE Transactions on Consumer Electronics, Vol. 60, No. 4, November 2014

[16] J. Yagnik, D. Strelow, D. A. Ross, and R. S. Lin, “The power of A. E. Çetin (F’09) studied Electrical Engineering at Middle
comparative reasoning,” in Proc. IEEE International Conference on East Technical University. After getting his [Link]. degree, he
Computer Vision, Barcelona, Spain, pp. 2431-2438, Nov. 2011. got his M.S.E and Ph.D. degrees in Systems Engineering from
[17] T. Dean, M. A. Ruzon, M. Segal, J. Shlens, S. Vijayanarasimhan, and J. the Moore School of Electrical Engineering at the University
Yagnik, “Fast, accurate detection of 100,000 object classes on a single of Pennsylvania in Philadelphia. Between 1987-1989, he was
machine,” in Proc. IEEE Conference on Computer Vision and Pattern Assistant Professor of Electrical Engineering at the University
Recognition, Columbus, USA, pp. 1814-1821, June 2013. of Toronto, Canada. Since then he has been with Bilkent
[18] R. Khan, A. Hanbury, J. Stöttinger, and A. Bais, “Color based skin University, Ankara, Turkey. Currently he is a full professor. He spent 1994-1995
classification,” Pattern Recognition Letters, vol. 33, no.2, pp. 157-163, academic year at Koç University in Istanbul, and 1996-1997 academic year at the
Jan. 2012. University of Minnesota, Minneapolis, USA as a visiting associate professor. He is
[19] R. L. Graham and F. F. Yao, “Finding the convex hull of a simple involved in Multimedia Understanding Through Semantics and Computational
polygon,” Journal of Algorithms, vol. 4, no. 4, pp. 324-331, Dec. 1983. Learning Network of Excellence research MUSCLE-ERCIM, computer vision
based wild-fire detection research VBI Lab, biomedical signal and image
processing research , MIRACLE project , signal processing research for food
BIOGRAPHIES safety and quality applications, wavelet theory, inverse problems and used to carry
out research related to Turkish Language and Speech Processing. He is currently a
F. Erden received his [Link] and M.S degrees in Electrical member of the SPTM technical committee of the IEEE Signal Processing Society.
and Electronics Engineering from Bilkent University, He founded the Turkish Chapter of the IEEE Signal Processing Society in 1991.
Ankara, Turkey. Since 2009, he has been a Ph.D. Student in He was on the editorial boards of EURASIP Journals, Signal Processing and
the Department of Electrical and Electronics Engineering at Journal of Advances in Signal Processing (JASP). Currently, he is the Editor-in-
Hacettepe University, Ankara, Turkey. His research Chief of Signal, Image and Video Processing SIViP, and a member of the editorial
interests include sensor signal processing, sensor fusion, boards of IEEE CAS for Video Technology and IEEE Signal Processing
multimodal surveillance systems and object tracking. Magazine. He holds four US patents.

Authorized licensed use limited to: GMR Institute of Technology. Downloaded on January 24,2025 at 08:50:57 UTC from IEEE Xplore. Restrictions apply.