2022 2nd International Conference on Advance Computing and Innovative Technologies in Engineering (ICACITE)

IoT and Blockchain Based Intelligence Security

System for Human Detection using an Improved
ACO and Heap Algorithm
2022 2nd International Conference on Advance Computing and Innovative Technologies in Engineering (ICACITE) | 978-1-6654-3789-9/22/$31.00 ©2022 IEEE | DOI: 10.1109/ICACITE53722.2022.9823827

Md. Akkas Ali Dr. B. Balamurugan Dr. Vandana Sharma

Ph.D Scholar, School of Computing Professor, School of Computing Professor, School of Computing
Science and Engineering Science and Engineering Science and Engineering
Galgotias University Galgotias University Galgotias University
Greater Noida, U.P, India Greater Noida, U.P, India Greater Noida, U.P, India
[email protected]
md.21scse3010038@galgotiasuniversi u.in
ty.edu.in
Abstract— In the modern processed world, it becomes more topic, as well as creating parallel deployments that can take
necessary to certify humans in a very secure way. There are advantage of new parallel software and hardware. Regarding
modern square measure applications such as online banking or image processing, ACO used only for edge detection and edge
online search usage techniques that are depended on passwords, linking till now. But we shall try to modify ACO algorithm
keys or individual identification card. These technologies and according to degree constraint local Search and use this
processes carry the danger that information may be forgotten, improved version of ACO for enhancing, Binarizing,
lost, or perhaps stolen. Therefore, the forms of identification Thinning, Minutiae tracking and clear false minutiae.
promise a singular thanks to being ready to certify humans. A
secure and confidential identification technique is the use of II. BACKGROUND
fingerprints. We proposed here IoT and Blockchain Based
Intelligence Security System for Human Detection using an Till now researchers used ACO algorithm for solving the
Improved ACO and Heap Algorithm. We proposed IoT and constrained base MST problems. For example, capacitated
Blockchain Based Technology for ensuring the security of our MST and generalized MST problems and in [1]-[4], in order
system. to detect edges, they have used ACO Algorithm. In [5]-[6],
histogram equalization techniques are proposed for image
Index Terms— ACO Algorithm, IoT and Blockchain, Heap enhancement and medical image enhancement respectively.
Algorithm, Human Recognition, Human Detection. Agarwal, D., Bansal, A. [7] proposed Assessment of Latent
Fingerprint Image quality based on level one and level two. In
I. INTRODUCTION [8] proposed Automatic Latent Fingerprint Identification
Solving the identification and detection problems of System. Praseetha, V.M.,S. Vadivel.[9] proposed enrolment
human beings can be classified by two different types with and matching of Fingerprints using Minutiae Tree. Lu,
various dimensions of identification and verification. The Guanlin, et al. [10] proposed a technique for texture
verification process defines to the issues of proving or enhancement. Joshi, S [11] Discrete Wavelet Transform
disproving a human’s promising identification. Reliability of Based Approach for Touchless Fingerprint Recognition. Wu,
human detection is critical in various day-to-day operations. Aihua, et al. [12] proposed an algorithm on topology
For example, managing access to physical objects and consistency for image matching. Friesnitz, J., Rathgeb, C.,
privileges on laptops or IoT devices is becoming increasingly Buchmann, N. et al [13] proposed touchless 2D fingerprint
important to prevent abuse. There is a growing interest in recognition.
cheap and reliable human detection in various civils,
commercial and financial applications. An automated process III. MOTIVATION
of verifying the similarities between two or more human’s IoT and Blockchain based mostly security, like human
fingerprint is known as automated fingerprint recognition. detection exploitation fingerprint authentication, is proved to
Because of uniqueness and persistence over time, fingerprints be each safer and convenient than passwords, creating
have been used for over a century and have more recently been fingerprint sensing associate in nursing progressively
automated i.e., biometrically thanks to advances in computing common and product-differentiating feature in sensitive
power. Due to the capable of obtaining from various available phones, tablets and PCs. However, fingerprint authentication
sources and establishment uses and collection through proper additionally raises security issues that may best be addressed
authentication sources like police as well as immigration also. with protections purposeful for bioscience. Junction helps
Various minutiae can be found in human finger, among them guarantee biometric information protection through the watch
only two minutiae are mostly used, one is called ridge ending purpose security suite of options and architectures that
and another called ridge bifurcation and left are cleared by accommodate the complete vary of market wants. So, for a
false removal stage. The fingerprint recognition problem is higher security we want fingerprint-based authentication
sorted into 3 sub-domains: fingerprint registration, system. we want to stay faster and higher performance from
verification, and fingerprint identification. the system. The performance depends upon the standard of
pictures and also the size of example. So, our main work is
In terms of applications, using ACO (Ant Colony
capturing the fingerprint image and build a higher quality of
Optimization) to solve dynamic, multi-objective, stochastic,
continuous, and mixed-variable optimization issues is a hot

978-1-6654-3789-9/22/$31.00 ©2022 IEEE 1792

Authorized licensed use limited to: Bahria University. Downloaded on April 05,2024 at 14:25:28 UTC from IEEE Xplore. Restrictions apply.
 2022 2nd International Conference on Advance Computing and Innovative Technologies in Engineering (ICACITE)

fingerprint example and it born-again into a computer memory E. Software Level Design
unit stream and detain an information. Software level design can be included minutiae
IV. OBJECTIVES AND SPECIFIC AIMS extraction, template construction, storing the template or
processing data as byte streams to decentralized Blockchain
The main objectives and specific aims are: database system and fingerprint matcher.
x To study and find the limitations of Basic ACO
Algorithm.
x To develop the improved version of Basic ACO
Algorithm by removing the limitations.
x To apply the improved version of Basic ACO
Algorithm in the required steps of our IoT and
Blockchain based Intelligence Security System.
x To increase the image quality.
x To keep data in database (Decentralized Blockchain)
as byte stream.
x To establish a better security system.
x To develop an IoT and Blockchain based Intelligence
Security System.
x To get a better accuracy (Recognition Rate).
x To detect the Human (Male or Female).
V. METHODOLOGY
A. System Implementation
Our IoT and Blockchain based intelligence security
system for human detection using an improved version of
ACO algorithm and Heap algorithm consist of three parts as
shown in fig.1.
x To improve the ACO algorithm
x To detect human using this improved ACO algorithm
and Heap algorithm
x To store the data in decentralized Blockchain as byte
streams for ensuring the security of our system using
Secure Hash Algorithm (SHA).
B. Improving ACO Algorithm
Graph based basic ACO works as without local search
and without degree constraint. We have proposed to modify
the basic ACO algorithm with local search and with degree
constraint. This improved ACO algorithm can be used to Fig. 1. Proposed Methodology
represent the fingerprint data by detecting the core point of
fingerprint which can be the root of Heap tree and minutiae
i.e., bifurcation and end point can be the child node of Heap
tree.
C. Detecting Human
We have proposed here an IoT and Blockchain based
Intelligence Security System for Human Detection by
fingerprint recognition using an Improved ACO and Heap
Algorithm. We proposed IoT and Blockchain Based
Technology for ensuring the security of our system.
D. Hardware Level Design: Sensor Fig. 2. Fingerprint Sensor (IoT device)[13]
Fingerprint image can be live scanned or captured by an
electronic device called scanner or sensor device as shown in F. Minutiae Extractor
fig. 2 . For the development of a minutia extractor, we can
process the pre-processing, minutia extraction and post-
processing stages.

1793
Authorized licensed use limited to: Bahria University. Downloaded on April 05,2024 at 14:25:28 UTC from IEEE Xplore. Restrictions apply.
 2022 2nd International Conference on Advance Computing and Innovative Technologies in Engineering (ICACITE)

G. Pre-processing
In pre-processing stage, we can do two vital operations
such as fingerprint image enhancement and binarization. Then
next step is binarization by using ACO algorithm.
H. Image Enhancement
We can enhance the inputting fingerprint image by using
us modified ACO algorithm for increasing the image quality.
It is very necessary to get accurate result from final stage.
I. Image Binarization Fig. 3. Minimum heap tree from minutiae points

Binarization is the method that converts to a binary image

from gray image. This process is used to increase the different
between ridges and valleys. The pixels of black color reflect
Fingerprint
ridges and the pixels of white color reflect the valleys. it
converts the 8-bit gray image to a 1 bit for ridge and a 0 bit for Information Database
valley. We can do these works using improved version of (Template as (Decentralize
ACO algorithm. Byte Stream) d Blockchain)

J. Minutiae extraction
The minutiae extraction can be done by following two
steps as the thinning process and minutiae marking process. Fig. 4. Template storing process in our system

K. Thinning Minimum Heap represents the total information of

The darkening of all ridge lines can be reduced by thinning minutiae. We keep this information in database
the image. After thinning process, original location remains (Decentralized-Blockchain) as byte stream for verification
and identification.
same and destruct the extra pixel of ridges to a width of just
one pixel. Thinning process can be based on the improved R. Fingerprint Matching
version of the ACO. We can develop a system for matching fingerprint using
L. Minutiae Marking heap algorithm.
Minutiae marking is the most important step. We can S. Accuracy Calculation
mark the minutiae using the modified ACO algorithm. A hit value can be calculated by applying the above-
M. Post-Processing mentioned algorithm that provides the number of matching
minutiae by which an accuracy can be calculated. The
The post-processing includes the false minutiae removal. equation is defined as following:
N. Remove False Minutiae Accuracy =Hit/N;
To get best accuracy, we need to remove false minutiae.
We can use the improved version of the ACO algorithm to N=Number of nodes whichever is minimum.
remove false minutiae. VI. RESULTS ANALYSIS
O. Template Construction The experiment can be performed by using Java, Matlab
We have proposed to create the template as a byte stream and tested on databases MySQL. Finally, we get result from
from fingerprint as shown in fig. 4 and store more secure way pre-processing stage, matching stage, identification (one to
in decentralized Blockchain database system. many) matching, verification system (one to one) matching,
accuracy and detection as human is male or female.
P. Getting information from Fingerprint using Heap
Algorithm A. False Rejection Rate (FRR)
The concept of property of heap tree can be used to gather When an authorized user tries to access the biometrics
information from fingerprint. We use the heap algorithm to system, falsely he/she is rejected. FRR is practically used to
gather the information from fingerprint. measure these types of incidents. An FRR system practically
establishes dividing the false rejection by number of nodes.
Q. Constructing Heap Tree using Heap Algorithm
To construct the heap tree using heap algorithm, we detect (%) FRR = (FR / N) .100
the minutiae and calculate the distance (d) between two Where FR= total falsely missing nodes. N = Number of
minutiae points. Minimum heap is constructed from d where nodes.
N1, N2, … Nn are the nodes. Nodes are representing the B. False Acceptance Rate (FAR)
minutia points of fingerprint image as shown in fig. 3.
When an unauthorized user tries to access the biometrics
system, falsely he/she is accepted. FAR is practically used to
measure these types of incidents. An FAR system practically
dividing the false acceptance by number of nodes.
(%) FAR=(FA / N).100
Where FA means total falsely accepted nodes, N= Number

1794
Authorized licensed use limited to: Bahria University. Downloaded on April 05,2024 at 14:25:28 UTC from IEEE Xplore. Restrictions apply.
 2022 2nd International Conference on Advance Computing and Innovative Technologies in Engineering (ICACITE)

of nodes. we can use ridge density or amount of amino acid. For

ensuring the security issues, faster, availability, scalability
C. Byte Stream
and integrity, we can use decentralized Blockchain based
Byte string is better than blob for fingerprint database storage.
authentication system. We shall keep data as byte string after
hashing into the decentralized Blockchain system for securing VIII. CONCLUSION
our data and also providing availability and integrity. Heap-based fingerprint recognition and human detection
TABLE I. COMPARISON OF VARIOUS DATA TYPES
algorithms can be superior and that are simpler but stronger
than orientation-based fingerprint matching algorithms. The
Data type Size work can be carried out sequentially. Pre-processing, feature
Blob 64k extractor, post-processing, template construction, storing the
Byte stream 4k template as byte stream to the IoT and Blockchain based
decentralized database, fingerprint matcher for finding the
matching or not matching. Finally, we can take decision as
Blob’s max size is 64Kb where byte string is 4Kb. If
user is recognition or not. If recognition can be found, then the
fingerprint database size is 64 Kb and we use byte string then
user is male or female can be detected by calculating the ridge
we can keep 14 to 15 templates in database where blob’s size
density or amount of amino acid founded in their fingerprint,
is 64 Kb. So, for a small template, a byte string is better
The real-life fingerprint data can be collected from users and
solution.
kept it in database as decentralized Blockchain after hashing
D. Identification (One to Many) by using Secure Hash Algorithm (SHA) in order to secure our
The comparison of biometric information of the user with process.
the template data of several users, which are taken over in the
during the registration is known identification. The REFERENCES
identification process is also known as one-to-many matching. [1] Md. Akkas Ali, “Improved and efficient Ant Colony Optimization
The result can be whether or not the user's fingerprint is algorithms for construction Constraint Based Minimum Degree
identified. Spanning Tree (ACO-CBMDST) of unweighted connected
graph,”M.Sc dissertation, Dept. of CS, Azteca University, Mexico,
E. Verification (One to One) 2021.
[2] Sergio Baltierra, Jonathan Valdebenito, Marco Mora, A proposal of
The comparison of biometric information of the user with edge detection in images with multiplicative noise using the Ant
the template data of one user, which are taken over in the Colony System algorithm, Engineering Applications of Artificial
during the registration is known verification. The verification Intelligence,Volume 110, ISSN 0952-1976, 104715, 2022.
process is known as one-to-one matching. The result can be [3] Fidanova, S, ACO for Image Edge Detection. In: Ant Colony
found or not. Optimization and Applications. Studies in Computational Intelligence,
vol 947. Springer, Cham, 2021.
F. Detection [4] Gulpi Qorik Oktagalu Pratamasunu, Olief Ilmandira Ratu Farisi, A
We can detect human as he / she is X or Y and also male Neutrosophic Approach to Edge Detection Using Ant Colony
Optimization, International Journal on Electrical Engineering and
or female. Informatics - Volume 13, Number 3, September 2021.
G. Comparison of Accuracy [5] Vijayalakshmi, D., Nath, M.K. A Novel Contrast Enhancement
Technique using Gradient-Based Joint Histogram
The proposed system can be compared with previous Equalization. Circuits Syst Signal Process 40, 3929–3967, 2021.
finger-print recognition system and the result as a comparison [6] Upendra Kumar Acharya, Sandeep Kumar, Genetic algorithm based
study as like shown in table II. adaptive histogram equalization (GAAHE) technique for medical
image enhancement, Optik,Volume 230, 166273, 2021.
TABLE II. TABLE II. COMPARISON OF ACCURACY [7] Agarwal, D., Bansal, A., Assessment of Latent Fingerprint Image
Quality Based on Level 1, Level 2, and Texture Features, International
Methods Accuracy (%) Year Conference on Innovative Computing and Communications. Advances
[8] 98.98 2022 in Intelligent Systems and Computing, vol 1165. Springer, 2021.
[9] 92.5 2011 [8] Uttam U Deshpande, V. S. Malemath, A Study on Automatic Latent
Fingerprint Identification System, Journal of Computer Science
[13] 98.78 2021 Research, Volume 04, Issue 01, January, 2022.
Proposed Method Higher than above and can 2022 [9] Praseetha, V. M., and S. Vadivel. "Enrolment and Matching of
be detected human as Fingerprints using Minutiae Tree.", Journal of Computer Science, 15
he/she is X or Y and also (3): 357.371, 2019.
male or female. [10] Lu, Guanlin, et al. "MSCNet: A Framework with a Texture
Enhancement Mechanism and Feature Aggregation for Crack
VII. DISCUSSION Detection." IEEE Access, 2022.
[11] Joshi, S, Discrete Wavelet Transform Based Approach for Touchless
We have proposed here an IoT and Blockchain Based Fingerprint Recognition, Proceedings of International Conference on
Intelligence Security System for Human Detection using an Data Science and Applications, Springer, Singapore, vol 288, 2022.
Improved ACO and Heap Algorithm. We can use modified [12] Wu, Aihua, et al. "Image Matching Algorithm Based on Topology
ACO and Heap algorithms for recognition and for detection, Consistency of Bidirectional Optimal Matching Point Pairs." Sensors
and Materials 34.2, 493-514, 2022.
[13] Friesnitz, J., Rathgeb, C., Buchmann, N. et al. An overview of
touchless 2D fingerprint recognition. J Image Video Proc, 8 2021.

1795
Authorized licensed use limited to: Bahria University. Downloaded on April 05,2024 at 14:25:28 UTC from IEEE Xplore. Restrictions apply.