Driving Aid for Rotator Cuff Injured Patients using
Hand Gesture Recognition
KRISHNASREE VASAGIRI
Department of Electronics and Communications Engineering,
VNR Vignana Jyothi Institute of Engineering and Technology, Hyderabad,
INDIA
Abstract: - Gesture recognition is a way for computers to understand how humans move and express
themselves without using traditional methods like typing or clicking. Instead of relying on text or graphics,
gesture recognition focuses on reading body movements, such as those made by the hands or face. Currently,
there is a specific interest in recognizing hand gestures by analyzing the veins on the back of the hand.
Scientists have found that each person has a unique arrangement of veins beneath the skin of their hand. When
the hand moves, the position of these veins changes, and this change is considered a gesture. These gestures are
then translated into specific actions or tasks by coding the hand movements. This technology is particularly
helpful for individuals with rotator cuff injuries. The rotator cuff is a group of muscles and tendons in the
shoulder that can get injured, causing pain and limiting movement. People with these injuries may have
difficulty steering a car, especially if their job or sport involves repetitive overhead motions. With gesture
recognition technology, a person can control the car by simply moving their wrist, eliminating the need to use
the shoulder. In summary, gesture recognition technology reads the unique patterns of hand veins to interpret
hand movements, making it a practical solution for individuals with rotator cuff injuries who may struggle with
certain tasks, like steering a car.
Key-Words: - Rotator cuff, hand gestures, Complex Walsh transform, Sectorization, Dorsal hand vein
arrangements, Kalman filter, Discrete wavelets.
1 Introduction
A gesture is a way of showing actions or emotions
visually, like using body and hand movements.
There are two types of gestures: passive ones,
where the body or hand position gives a clue, and
active ones, where the body or hand changes to
convey information. Gestures help people interact
with computers in a different way than using
regular tools. Instead of relying on hardware like
keyboards or mice, computers can understand
human intent by recognizing body movements or
changes in limb positions. Scientists have been
working on improving the technology to recognize
hand gestures, [1], [2], [3], and it has various
applications such as understanding sign language
and enhancing experiences in virtual reality.
1.1 Rotator Cuff Injury
The rotator cuff as shown in Figure 1 is like a
bunch of muscles and tendons around your
shoulder joint. It keeps the upper arm bone in place
in the shoulder. If your shoulder hurts, it might be
because of a rotator cuff injury. The pain can get
worse if you sleep on the sore shoulder. People
who do jobs or play sports that involve a lot of
overhead movements, like carpenters, painters,
tennis players, and baseball players, are more likely
to get rotator cuff injuries. The risk of these injuries
goes up as you get older. The injury could be a
simple inflammation or a complete tear, [4], [5].
Having a rotator cuff problem is painful and
limits your ability to move your shoulder. A big
injury to the shoulder or the slow wearing down of
tendon tissue can cause rotator cuff problems.
Doing a lot of overhead movements over a long
time, developing bone spurs around the shoulder, or
continuous upward actions can make the tendon
worse.
In these situations, people with rotator cuff
injuries may find it hard to put pressure on their
shoulders, like when driving a car. If a driver has a
rotator cuff injury, he/she might struggle to turn the
steering wheel and may even do accident because
he/she can't control the car properly due to pain.
However, if we design an aiding tool for people
suffering from rotary cuff injuries which allows the
driver to control the car by just moving the hands
without putting pressure on shoulder. This way,
Received: March 11, 2023. Revised: March 12, 2024. Accepted: April 14, 2024. Published: May 13, 2024.
WSEAS TRANSACTIONS on SIGNAL PROCESSING
DOI: 10.37394/232014.2024.20.3
Krishnasree Vasagiri
E-ISSN: 2224-3488
20
Volume 20, 2024
they can steer the car effectively with less physical
effort and better control.
Fig. 1: Rotary cuff diagram
1.2 Dorsal Hand Veins
The dorsal hand vein is the arrangement of the
blood veins beneath the dermis of the hand. This
pattern is exclusive to every person and cannot be
tampered with or forged by anybody. Due to this
sturdy nature of characteristics, this is used as a
Biometrics in recent days. This arrangement is used
as the representation of the hand gesture and
utilized for detecting the rotation of the hand, [6],
[7].
1.3 Detection and Recognition of Hand
Gesture
The objective of this research is to develop a novel
technique to recognize the gestures of the hand. It
is proposed a completely new algorithm for the
detection of the angle of rotation through a hand
gesture recognition algorithm. The techniques or
methods of recognition of gestures of hand need to
be satisfied for various parameters, [8], [9]:
Flexibility: (can be applicable to any
requirements), [10].
Practicality: The approach shall be practically
precise enough to be used and perfectly detect the
required gestures.
Reliability it should be reliable
Robustness: The system should be robust such that
it can identify the gestures even though the light is
bad, the background is not clear and the position of
the object is rotated.
Scalability: It should be scalable.
Versatility: The algorithm should work on
different hands uniformly without any variation in
the performance even if the sizes of the hands are
different or the color of the hands is different.
1.4 Tracking of Dorsal Hand Veins
Dorsal hand veins are one such feature that is user
independent and satisfies all the conditions
mentioned. Hence, in our work, the orientation of
dorsal hand veins is detected and is used to steer
the car accordingly.
1.5 Related Work
Hand gesture recognition is carried out using
Neural networks and Perceptron convergence
algorithm provide noise rejection and more
processing speed, [11].
Maximum likelihood recognition, [12].
Kalman filter approach and Bayesian
framework, [13].
With Haar Classifier and adaboost Algorithm,
[14].
A Kalman filter and hand blobs analysis.
Markov models are used for gesture
recognition and are robust for ASL gestures
and background clusters and recognition
accuracy of up to 98% has been achieved, [15]
PCA, Pruning, and ANN, [16].
Via LAN & Wireless Hardware control using
mathematical algorithm, [17].
With Eigen dynamics analysis by modeling
hand dynamics with EDA and likelihood edge.
It combines histograms of color and features of
edges, [18].
Haar Cascade, Adaboost Algorithm, Hidden
Markov Model and Douglas Peucker
Algorithms achieved average recognition rate
of 91% on data set of 16 gestures, [19].
Kalman filter using a 3D depth sensor. They
identified the probable images of hands with
the help of clusters of motion, and predefined
wave motion, and tracked the positions of the
hand using the Kalman filter, [20].
Boundary lines of palms and vein patterns were
detected employing canny edge detection
techniques, [21].
Depth-based segmentation and Haarlets, [22].
Human computer interaction using hand
gesture with accuracy of 95.44%, [23], [24].
Dynamic time warping and Histogram of
oriented gradient features, [25].
2 Materials
To identify the rotation of the hand, we use images
of vein patterns beneath the skin on the upper part
of the hand. These vein patterns can't be seen or
captured with a regular camera that uses visible
light. Instead, we use NIR (Near-Infrared)
capturing technology with a special camera that has
a monochrome Near-Infrared CCD camera fitted
with an infrared lens.
WSEAS TRANSACTIONS on SIGNAL PROCESSING
DOI: 10.37394/232014.2024.20.3
Krishnasree Vasagiri
E-ISSN: 2224-3488
21
Volume 20, 2024
Two sources of infrared (IR) energy are
directed at the upper hand, and the blood's
hemoglobin absorbs this energy, making the vein
patterns visible as thick lines. A CCD camera then
captures the image, turns it into digital data, and
stores it in a database for later analysis. You can
see an example of how these vein patterns are
arranged in Figure 2 and Figure 3.
Fig. 2: Dorsal hand vein image-normal
Fig. 3: Dorsal hand vein image-Near infrared
3 Method
In Figure 4, you can see the steps for tracing hand
vein gestures using Kalman filtering. Figure 5
shows the algorithm for recognizing vein gestures,
using methods like Histogram of Oriented
Gradients, Reversible Complex Hadamard
Transform, and Discrete Wavelet Transform
(DWT). Here's a simple breakdown of what
happens:
Image Transformation:
The initial image is turned into a gray one, and a
Region of Interest (ROI) is selected.
Database Enhancement:
The database is made better by rotating input
images both clockwise and counterclockwise.
Various Analysis Methods:
Three different methods (Histogram of Oriented
Gradients, four-phase Reversible Hadamard
Transform, and DWT) are used for image analysis.
Feature Extraction:
Important features are taken from different parts of
the image and saved in the database for later
correlation to understand how the image is rotated.
Histogram of Oriented Gradients:
Gradients (changes in intensity) are calculated from
the image database, creating a feature database that
helps determine the rotated angle.
Reversible Complex Hadamard Transform:
Attributes are taken from quarters of coefficients of
the reverse Hadamard transform and stored in the
database for matching attributes and finding the
rotated angle.
Gestures Tracking with Kalman Filtering:
Kalman filtering, Sobel, and Canny edge detectors
are used to locate edges. The reference centroid is
calculated, and the Kalman filter helps track
coordinates.
In simple terms, these steps show how
computer programs can analyze images of hand
veins, understand the gestures, and figure out how
the hand is rotated using advanced techniques.
Fig. 4: Hand gestures tracing using Kalman
filtering
Fig. 5: The method of identifying vein patterns on
the back of the hands using techniques like DWT,
Histogram of Oriented Gradients, and Reversible
Input
seque
nce of
Image
s
Extraction of ROI,
Pre-processing,
Canny& Sobel
edge detection
Kalman
Filter
Calculatio
n of
Reference
(centroid)
Tracked
Output
Input
Image
Extraction
of ROI
and Pre-
processing
Discrete Wavelet
transform/Histogram
of oriented
gradients/Reversible
complex Hadamard
transform
Feature
vector
matching
Sectorization/Feat
ure vector
extraction
Data base
Output
WSEAS TRANSACTIONS on SIGNAL PROCESSING
DOI: 10.37394/232014.2024.20.3
Krishnasree Vasagiri
E-ISSN: 2224-3488
22
Volume 20, 2024
Complex Hadamard Transform is outlined in the
algorithm
The features taken from a given image are then
compared with those in the image database using
Euclidean distance to find the best match. This
matching process considers parameters such as
standard deviation, arithmetic average, skewness,
and median.
3.1 Pre-processing
To make the analysis more effective, we first cut
out the part of the image where the vein
arrangement is visible by cropping. We then
remove the rest of the image, and apply adaptive
histogram equalization to improve the image
quality by making the pixel values more even .
Once this processing is done, we save these
improved images in a database for training
purposes.
3.2 Discrete Wavelet Transform
The process of breaking down the Discrete Wavelet
Transform (DWT) into a pyramid structure
involves using low-pass and high pass filters on the
original image. This produces four second-level
coefficients: a smoothened image known as a low-
low (LL) partitioned image, which holds the
approximate details of the original image, and three
detailed partitioned images that show the horizontal
(LH), vertical (HL), and diagonal (HH) directions
of the original image. First level decomposition is
given in Figure 6 and second level decomposition
is given in Figure 7.
Fig. 6: First level DWT decomposition
Fig. 7: Second level DWT decomposition
The decision on which class an image belongs
to is made by measuring the Euclidean distance
between the features of the query image and those
of the training dataset.
Various types of wave filters are part of
families like Daubechies, Coiflets, Symlets, and
Fejer-Korovkin. In this case, Haar wavelet from the
Daubechies family is used. Haar Wavelet is a
combination of rescaled "square-shaped" functions
called “basis” functions, each with different scales.
Using Haar wavelets provides four coefficients of
Discrete Wavelet Transform (DWT):
approximation, horizontal, vertical, and diagonal.
These coefficients are important for extracting
features from the image.
3.3 Reversible Complex Hadamard
Transform
The complex Hadamard matrix H of order N, (and
H*complex conjugate transpose) is a unitary matrix
with symbols ±1, ± j, where j = 1, is shown in
equation (1)
HH* = H*H =
The matrix
is an illustration of
a 2nd order complex Hadamard matrix. Complex,
[26], [27] of The Hadamard higher order matrices
are computed coercively by using Kronecker
product is given in equation (2)
(2)
The direct and inverse reversible complex
Hadamard matrices are given by equation (3), (4).
Basic 2x2 matrix is given by equation (5)
(3)
Where,
(5)
The direct and forward reversible complex
Hadamard matrices of eighth order are given in
equations (6) and (7).
(6)
WSEAS TRANSACTIONS on SIGNAL PROCESSING
DOI: 10.37394/232014.2024.20.3
Krishnasree Vasagiri
E-ISSN: 2224-3488
23
Volume 20, 2024
(7)
3.4 Sectorization
The complex Walsh transform consists of real and
imaginary parts that are divided into sections with
proportions of 4, 8, and 12. Initially, the Walsh
transform is broken down into 4 regions. To further
break down these regions into 8 and 12, the 4-
section transform is treated as the foundation, and
the decomposition continues. This process is
illustrated in Table 1, [28]. A set of attributes is
then created by averaging all Walsh coefficients for
each region. This value is unique for each image
since the distribution of sequences for every hand
image is distinct in different regions.
3.5 Feature Extraction
A set of features is created for each quadrant,
which includes values like average, median,
standard deviation, skewness, and median. Since
each arrangement of veins is unique, the sequence
distribution patterns and feature vectors are also
unique. The number of vectors related to features is
minimal compared to the significant Walsh
coefficients that carry the most signal energy. This
greatly reduces complexity and execution time. The
division of Walsh transform coefficients into
various sectors follows a standard procedure, as
indicated in Table 1.
Table 1. Sector’s division
Sign of
Sal
Sign of
Cal
phase
Quadrant
Assigned
Sector
positive
positive
00-900
1st
1
positive
negative
900-1800
2nd
2
negative
negative
1800-2700
3rd
3
negative
positive
2700-3600
4th
4
3.6 Veins Tracking using a Kalman Filter
The Kalman filter is a commonly used tool for
tracking the movement of various objects. It makes
predictions about the object's state by considering
results influenced by noise and measurements. If
possible, it assesses the actual current state of the
object. Kalman filters find application in different
linear dynamic systems. In this context, a "state"
refers to any assessable physical parameter, such as
the object's location, temperature, velocity, voltage,
and so on. The Kalman filter operates in two stages
and is used in a recursive manner. After each
iteration, it predicts and updates the status of the
parameter.
The prediction is based on the current location
of the moving object, using previous information.
After measuring the object's latest position
(denoted as Zt), it combines this with the guessed
current location (Xt) to obtain a more accurate
estimate of the object's current location, Xt. The
equations governing the operation of the Kalman
filter are presented in equations (8) to (14).
At the level of prediction: Predicted (a priori) state
(8)
Predicted (a priori) estimate covariance
(9)
Level of updation: measurement residual:
(10)
Residual covariance
(11)
Optimized gain:
Updated (a posteriori) state estimate
(13)
Updated (a posteriori) estimate covariance
(14)
is ongoing state vector at time t from the
estimation of Kalman filtering, is the
assessment vector computed at time t,
calculates the assessed exactness of at time t, F
characterizes the movement of the from one state to
the other, that is projection of earlier state vector is
to the further, presuming that noise is absent. (e.g.
no acceleration). H defines the mapping from the
state vector, to the measurement vector, .
Gaussian process is defined as Q and assessment
noise is defined as R. These 2 parameters describe
the system’s variance. The control-input variables
are B, u.
3.7 Orientation of Histogram
The histogram of oriented gradients is a way to
describe an image by looking at the appearance and
shape of small neighborhoods within it. It focuses
on the spread of intensity gradients or the directions
WSEAS TRANSACTIONS on SIGNAL PROCESSING
DOI: 10.37394/232014.2024.20.3
Krishnasree Vasagiri
E-ISSN: 2224-3488
24
Volume 20, 2024
of edges in the image. Initially designed to detect
human faces, we are suggesting its use in our work
to identify the orientation of veins on the back of
the hand, [29].
4 Results and Discussions
In Figure 8 and Figure 9, we can see images from
the database featuring hands rotated both clockwise
and anti-clockwise. Figure 10 shows the image of
the hand pattern, while Figure 11 displays the
cropped image and Figure 12 presents the
thresholded image. Results from Discrete Wavelet-
based hand vein tracking are demonstrated in
Figure 13 and Figure 14. For reverse Hadamard
transform-based hand vein tracking, you can refer
to Figure 15 and Figure 16. The input hand image
for Histogram of Oriented Gradients is illustrated in
Figure 17, and Figure 18 displays the angle rotated
using Histogram of Oriented Gradients. Results
from the Kalman filter using Sobel and Canny
Edge detection can be found in Figure 19, Figure
20, Figure 21, Figure 22, Figure 23 and Figure 24.
Fig. 8: Database of the hand rotated clockwise
Fig. 9: Database of the hand rotated anti-clockwise
Fig. 10: Input image
Fig. 11: Cropped Image
WSEAS TRANSACTIONS on SIGNAL PROCESSING
DOI: 10.37394/232014.2024.20.3
Krishnasree Vasagiri
E-ISSN: 2224-3488
25
Volume 20, 2024
4.1 Results of Discrete Wavelet Transform
Fig. 12: Thresholded image
Fig. 13: Test Image for DWT
Fig. 14: Displaying angle rotated using DWT
4.2 Results of Reverse Hadamard
Transform
WSEAS TRANSACTIONS on SIGNAL PROCESSING
DOI: 10.37394/232014.2024.20.3
Krishnasree Vasagiri
E-ISSN: 2224-3488
26
Volume 20, 2024
Fig. 15: Input image with 200 rotation of hand for the reverse Hadamard transform
Fig. 16: Displaying the angle 200 rotated using Reverse Hadamard transform
4.3 Results of Histogram of Oriented
Gradients
Fig. 17: Input hand for with 30 rotation of hand Histogram of Oriented Gradients
Fig. 18: Displaying the angle with 30 rotated using Histogram of Oriented Gradients
WSEAS TRANSACTIONS on SIGNAL PROCESSING
DOI: 10.37394/232014.2024.20.3
Krishnasree Vasagiri
E-ISSN: 2224-3488
27
Volume 20, 2024
4.4 Results of Kalman filter (using Sobel
and Canny edge detection)
Fig. 19: Input image for Edge detection
Fig. 20: Extracted veins using Sobel Edge detection
Fig. 21: Extracted Veins using Canny edge
detection
Fig. 22: Input sequences of images for Kalman
Filter
Fig. 23: Output of the coordinates of image sequence using Kalman Filter (Sobel)
Fig. 24: Output of the coordinates of image sequence using Kalman Filter (canny)
WSEAS TRANSACTIONS on SIGNAL PROCESSING
DOI: 10.37394/232014.2024.20.3
Krishnasree Vasagiri
E-ISSN: 2224-3488
28
Volume 20, 2024
5 Conclusions
We tested the algorithms with rotations from -60
degrees to +60 degrees and achieved successful
detection of rotations in both clockwise and anti-
clockwise directions. There are four methods we
used for tracking hand rotation: DWT, reverse
Hadamard, orientation histogram, and Kalman
filter. Each method produced different results, and
their accuracy varied. The details of the time it took
for each method are given in Table 2 for a
comparison of execution times.
Table 2. Comparison of execution times
Algorithm
DWT
Reversible
complex
Hadamard
transform
Orientation
of
Histogram
Time taken
for
execution
0.1126
msec
0.1043msec
1.1254
msec
We discovered that when tracking the
coordinates of the reference point with the Kalman
filter, the precision is 2 degrees. In terms of
execution time, we found that the Reversible four-
phase complex Hadamard transform is a better
method, taking less time to complete.
6 Future Scope
This technique will serve as a good initiation for
hand gesture based car driving. It will be very
much useful for driving wheelchair for the patients
who are suffering from paralysis. It can be used for
gaming incorporated with augmented reality using
hand gestures. It will serve as better biometric
technique with low cost and greater accuracy for
industrial and real time usage for authentication. It
can be used to develop remote control applications.
References:
[1] Noor Adnan Ibraheem, Rafiqul Zaman Khan,
“Survey on Various Gesture Recognition
Technologies and Techniques”, International
Journal of Computer Applications, Vol. 50 –
No.7, pp-38-44, 2012, doi: 10.5120/7786-
0883.
[2] Kawade Sonam P, V. S. Ubale, “Gesture
Recognition-A Review”, IOSR Journal of
Electronics and Communication
Engineering, Vol. 5, pp.19-26.
[3] Pavlovic, Rajeev Sharma, and Thomas S.
Huang “Visual interpretation of hand
gestures for human-computer interaction: A
review", IEEE Trans. Pattern Analysis and
Machine Intelligence, Vol. 19(7), pp.677 -
695, July 1997.
[4] R. H. Crusher, “Rotator cuff injuries”,
Accident and Emergency Nursing, Vol. 8,
pp-129 – 133, 2000.
[5] Frank W. Jobe, and Diane Radovich
Moynes, “Delineation of diagnostic criteria
and rehabilitation program for rotator cuff
injuries”, The American journal of sports
medicine, vol.10, No.6, pp-336-339, 1982,
doi: 10.1177/036354658201000602.
[6] Badawi. A. M, “Hand Vein Biometric
Verification Prototype: A Testing
Performance and Patterns Similarity” In
Proceedings of the 2006 International
Conference on Image Processing, Computer
Vision, and Pattern Recognition, IPCV'06,
Las Vegas, USA, PP- 3-9, June 26-29, 2006.
[7] Jain.A.K, Ross.A and Prabhakar.S, , “An
Introduction to biometric Recognition”, IEEE
Transactions on circuits and systems for
Video Technology, Vol 14, No 1, PP 4-20,
January 2004.
[8] Chang, V., Eniola, R.O., Golightly.L.,
Qianwen Ariel Xu “ An Exploration into
Human–Computer Interaction: Hand Gesture
Recognition Management in a Challenging
Environment”, SN Computer Science,
SCI. 4, 441, pp-1-17,2023.
[9] Oudah M, Al-Naji A, Chahl J., “Hand
Gesture Recognition Based on Computer
Vision: A Review of Techniques”, J.
Imaging, 6(8):73, pp-1-29, 2020 Jul 23, doi:
10.3390/jimaging6080073.
[10] Aashni Haria , Archanasri Subramanian , Niv
edhitha Asokkumar , Shristi Poddar , Jyothi
S Nayak, “Hand Gesture recognition for
Human Computer Interaction”, Procedia
computer science, Vol. 115, 2017, pp- 367-
374.
[11] Klimis Symeonidis, “Hand Gesture
Recognition Using Neural Networks,”
Degree of Master of Science in Multimedia
Signal Processing communications, School of
Electronic and Electrical Engineering, pp.1-
69, On August 23,2000.
[12] Attila Licsár, Tamás Szirányi, “Hand Gesture
Recognition in Camera-Projector System”,
Published in ECCV Workshop HCI, Prague,
Czech republic, May 16th, 2004, DOI:
10.1007/978-3-540-24837-8_9.
[13] Pragati Garg, Naveen Aggarwal and Sanjeev
Sofat, “Vision Based Hand Gesture
Recognition”, International Journal of
WSEAS TRANSACTIONS on SIGNAL PROCESSING
DOI: 10.37394/232014.2024.20.3
Krishnasree Vasagiri
E-ISSN: 2224-3488
29
Volume 20, 2024
Computer and Information Engineering, Vol:
3, No: 1, pp.186-191, 2009.
[14] Ruchi. M. Gurav ,Premanand K. Kadbe,
“Vision Based Hand Gesture Recognition
with Haar Classifier and AdaBoost
Algorithm”, International Journal of Latest
Trends in Engineering and Technology
(IJLTET),Vol. 5, Issue 2, pp- 155-160,
March 2015.
[15] Nguyen Dang Binh, Enokida Shuichi,
Toshiaki Ejima, “Real-Time Hand Tracking
and Gesture Recognition System”, GVIP 05
Conference, CICC, Cairo, Egypt, 19-21
December 2005.
[16] S.M. Shitole ,S.B. Patil ,S.P. Narote ,
“Dynamic Hand Gesture Recognition using
PCA, Pruning and ANN”, International
Journal of Computer Applications, Vol. 74,
No.2, pp.24-29, July 2013.
[17] Tarachand Saini,Savita Sivani, “ Real Time
Vision Hand Gesture Recognition Based
Media Control via LAN & Wireless
Hardware Control”, International Journal of
Modern Engineering Research (IJMER),
Vol. 3, Issue. 5, pp.3129-3133, Sep – Oct,
2013.
[18] Hanning Zhou and Thomas S. Huang,
“Tracking Articulated Hand Motion with
Eigen Dynamics Analysis “, Proceedings of
the Ninth IEEE International Conference on
Computer Vision (ICCV’03), Nice, France,
volume 2, pp- 1102-1109, 2003.
[19] Varsha Dixit, Anupam, “Real Time Hand
Detection & Tracking for Dynamic Gesture
Recognition”, Intelligent Systems and
Applications, 08, pp-38-44, 2015.
[20] Sangheon Park, Sunjin Yu , Joongrock Kim,
Sungjin Kim, Sangyoun Lee, “3D hand
tracking using Kalman filter in depth space”,
EURASIP Journal on Advances in Signal
Processing, pp.1-18, 2012,
https://doi.org/10.1186/1687-6180-2012-36.
[21] Mr. S. D. Raut, Dr. V. T. Humbe, “An
Approach to Boundary Extraction of Palm
Lines and Vein Pattern”, I.J. Image,
Graphics and Signal Processing, 12, pp 47-
52, 2014.
[22] Van den Bergh, Michael, “Combining RGB
and ToF cameras for real-time 3D hand
gesture interaction”, IEEE Workshop on
Applications of Computer Vision, 5-7, Kona,
HI, USA, January 2011.
[23] Ram Pratap Sharma and Gyanendra K.
Verma, “Human Computer Interaction using
Hand Gesture”, Science Direct, Eleventh
International Multi-Conference on
Information Processing, Procedia Computer
Science, Bangalore, India, 54, pp-721–727,
2015, DOI: 10.1016/j.procs.2015.06.085.
[24] Pathan RK, Biswas M, Yasmin S, Khandaker
MU, Salman M, Youssef AAF. Sign
language recognition using the fusion of
image and hand landmarks through multi-
headed convolutional neural network. Sci
Rep. 13(1):16975, 2023 Oct 9, doi:
10.1038/s41598-023-43852-x.
[25] Pat Jangyodsuk ,Christopher Conly, Vassilis
Athitsos , “Sign Language Recognition using
Dynamic Time Warping and Hand Shape
Distance Based on Histogram of Oriented
Gradient Features”, Proceeding of the 7th
international conference on pervasive
technologies related to assistive
environments, Rhodes, Greece, May 27-30,
2014.
[26] Jiasong Wu, Lu Wang, Lotfi Senhadji,
Huazhong Shu, “Sliding Conjugate
Symmetric Sequency-Ordered Complex
Hadamard Transform: Fast Algorithm and
Applications”, IEEE Transactions on
Circuits and Systems Part 1 Fundamental
Theory and Applications, Institute of
Electrical and Electronics Engineers (IEEE),
59 (6), pp.1321-1334, 2012.
[27] G. Rama Murthy and D. Praveen, “Complex-
valued Neural Associative Memory on the
Complex hypercube,” IEEE Conference on
Cybernetics and Intelligent Systems,
Singapore, 1-3 December 2004.
[28] Dr.Vinayak Ashok Bharadi, “Texture Feature
Extraction For Biometric Authentication
using Partitioned Complex Planes in
Transform Domain”, International journal of
advanced computer science and applications
(IJACSA), pp.39-46, 2012.
[29] Dalal, N. and Triggs B., “Histograms of
Oriented Gradients for Human Detection”,
Computer Vision and Pattern Recognition.
Proceedings of IEEE Computer Society
Conference San Diego, CA, USA (CVPR), 1,
pp.886-893, June 2005.
WSEAS TRANSACTIONS on SIGNAL PROCESSING
DOI: 10.37394/232014.2024.20.3
Krishnasree Vasagiri
E-ISSN: 2224-3488
30
Volume 20, 2024
Contribution of Individual Authors to the
Creation of a Scientific Article (Ghostwriting
Policy)
The authors equally contributed in the present
research, at all stages from the formulation of the
problem to the final findings and solution.
Sources of Funding for Research Presented in a
Scientific Article or Scientific Article Itself
No funding was received for conducting this study.
Conflict of Interest
The authors have no conflicts of interest to declare.
Creative Commons Attribution License 4.0
(Attribution 4.0 International, CC BY 4.0)
This article is published under the terms of the
Creative Commons Attribution License 4.0
https://creativecommons.org/licenses/by/4.0/deed.e
n_US
WSEAS TRANSACTIONS on SIGNAL PROCESSING
DOI: 10.37394/232014.2024.20.3
Krishnasree Vasagiri
E-ISSN: 2224-3488
31
Volume 20, 2024
