ONLINE CRIME REPROTING SYSTEM

S.SUMATHI1
K.SUMAN 2
CSE Department
CSE Department
Jeppiaar Institute of Technology
Jeppiaar Institute of Technology
Sriperumbudur, India
Sriperumbudur, India
[email protected]
Abstract— The Online Crime Reporting System (OCRS) is an
innovative web-based platform designed to enhance public
populations, including those with limited knowledge of
safety and streamline law enforcement processes by providing a official languages. The advent of the digital age has placed
convenient and efficient way for citizens to report crimes. increased emphasis on secure and efficient election processes,
Traditional methods of crime reporting, such as in-person visits facilitating the rise of electronic voting systems (electronic
to police stations, often suffer from limitations such as limited voting). The objective of this study is to introduce a
operating hours, language barriers. In response to these breakthrough approach that not only improves the
challenges, OCRS offers a user-friendly and accessible solution completeness and accuracy of the voting procedure, but also
that empowers citizens to report crimes from the comfort of advances the integration of facial recognition techniques
their own homes or on-the-go using any internet-enabled
through the HAAR cascade classifier. This proposed
device. The system's Reporting and Analytics Module generates
reports and analytics on crime trends, patterns, and hotspots, electronic voting system represents a step forward in the
providing valuable insights for strategic planning, resource pursuit of robust, secure, and accessible elections. By
allocation, and crime prevention efforts. Overall, the Online leveraging the power of advanced facial recognition
Crime Reporting System represents a significant advancement algorithms, the system introduces an additional layer of
in public safety technology, leveraging the power of the internet authentication to ensure registered voters can cast their
and digital communication to empower citizens, enhance votes. Maximum security. By using the HAAR cascade
collaboration between the community and law enforcement classifier architecture, carefully trained on a large dataset of
agencies, and ultimately contribute to safer and more secure facial images, the system achieves accurate voter
communities.
identification and verification in real-time. This acts as a
protection against potential threats such as voter fraud and
Keywords—User-centric, Geolocation, Encryption,
identity theft. The main benefit of this innovative system is
Administrative Dashboard, Incident Management, Analytics,
that it can optimize and improve the voting experience for
Mapping, Community Engagement, Proactive Crime
citizens. This eliminates the need for physical voter ID
Prevention.
documents and provides a seamless and convenient way for
voters to exercise their democratic rights.

I. INTRODUCTION II. MOTIVATION AND RELATED WORK

Electronic voting systems, commonly referred to as e- Electronic voting systems promise to transform democratic
voting, represent a technological advancement in the field of processes by making them more accessible, more
election processes, with the primary purpose of modernizing convenient, and more efficient, but implementing secure and
and optimizing the voting experience and addressing the reliable systems remains a challenge One proposed solution
various challenges associated with traditional paper-based involves facial recognition, leveraging the Haar cascade
voting systems is to deal with it. classifier to detect facial features. Although appealing, a
balanced assessment of its motivations, work involved, and
An electronic voting system consists of several major
challenges is important before considering widespread
components, including an electronic voting machine or
adoption. Facial recognition for voter authentication is not
online platform, a voter registration database, and a results
new. "A Secure e- Voting System Based on Biometric
tabulation system. These components work in harmony to
Authentication" (Patel et al., 2013) propose multimodal
streamline the voting process, making it more efficient,
approaches with fingerprints and faces, while "Secure E-
accurate, and accessible to more people. One of the main
Voting System Using Multimodal Biometric
benefits of electronic voting systems is that they can
Authentication" (Hassan et al., 2018) combines faces and
improve efficiency by reducing long lines and waiting times
irises. Additionally, Haar cascade classifiers, popularized by
at polling stations. Voters can vote quickly and
OpenCV, have achieved success in facial feature detection
conveniently, increasing overall turnout. Furthermore,
systems such as "Real-time Face Detection Using Haar
electronic voting systems greatly reduce the risk of human
Cascade Classifier" (Patil et al., 2018) and "Improved Face
error in vote tabulation and virtually eliminate problems
Detection Using Haar Cascade Classifiers" (Jadhav et
such as overvoting and under voting. This results in more
al2020). The main driver of using facial recognition in
accurate and reliable election results. Accessibility is
electronic voting is its potential to enhance voter security
another important aspect of electronic voting systems. They
and verification.This could provide an additional layer of
provide people with disabilities with alternative voting
protection against identity theft and fraud, potentially
methods and make it easier for them to participate in the
enhancing the integrity of elections.Additionally, it can
democratic process. Additionally, electronic voting systems
improve convenience and accessibility for voters,
often include features such as multilingual support to
suit different

XXX-X-XXXX-XXXX-X/XX/$XX.00 ©20XX IEEE

especially those living in remote areas or without proper approach to modernize and secure the electoral process. This
documentation.Finally, its potential expansion could system aims to enhance the accuracy, efficiency, and
theoretically appeal to more voters more effectively.Its integrity of voting by leveraging advanced technologies. In
appeal and significant challenges require careful this system, eligible voters' facial features are captured and
consideration.Privacy concerns related to the collection, stored securely in a database during the registration process.
storage, and use of facial recognition data are On the day of the election, voters can log in to the e-voting
paramount.Strong data protections and user consent are platform using their unique facial biometrics, eliminating
essential to mitigate these concerns.Additionally, the the need for physical identification documents. The system
accuracy and robustness of facial recognition systems are utilizes HAAR Cascade Classifiers to analyze and verify the
critical to ensuring voter trust and system reliability.They authenticity of the captured faces and verify OTP number,
may be prone to making mistakes in difficult conditions, ensuring that only legitimate voters can cast their ballots. By
obscuring their faces, or intentionally deceiving.Deploying incorporating face recognition and OTP verification, the
such systems nationwide will require significant investments system can prevent fraudulent voting attempts, protect
in technology infrastructure, training and public awareness against duplicate voting, and streamline the authentication
campaigns.These investments must be carefully weighed process, offering a convenient and user-friendly voting
against the potential benefits and feasibility of responsible experience for citizens. Additionally, voter confidence is
implementation.A clear legal and regulatory framework is bolstered, as advanced technology provides an extra layer of
essential to govern the use of facial recognition in electronic security, making it harder for unauthorized individuals to
voting, addressing concerns around privacy, security, interfere with the voting process. However, the successful
accessibility and discrimination of data.Without such implementation of this system requires addressing privacy
frameworks, the risk of abuse and discrimination becomes a concerns, ensuring transparency, and conducting thorough
significant risk.While electronic voting with facial testing to maintain its effectiveness and trustworthiness.
recognition offers exciting possibilities, it is imperative to
recognize and address the ethical, technological and legal
challenges it presents.Responsible implementation is based
on comprehensive research and development and a
commitment to addressing privacy concerns, ensuring
accuracy and comprehensiveness, and establishing strong
regulatory frameworks.Only then can we consider
widespread adoption of this technology in electronic voting
systems.

EXISTING SYSTEM

Before any election, eligible citizens must register

as voters. The Election Commission of India (ECI) oversees
the registration process and individuals can register online
or through designated voter registration centers. Once
registered, the names of eligible voters will be recorded in
the electoral roll, which is regularly updated by the ECI.
After successful registration, voters will receive a voter
identification card, also known as a voter photo
identification card (EPIC). This card serves as identification
during the voting process. Election Notification: When
elections are scheduled, the ECI will announce the dates,
constituencies, and polling stations for voting. On election
day, voting booths are located at different locations in each
constituency to facilitate voting. Voters arrive at their
designated polling place and present their voter ID card or
other approved identification to election officials. They
receive a ballot or use an electronic voting machine (EVM)
to vote. Electronic voting machines (EVMs)are widely used
in India for voting. Voters press the button next to the
symbol of the candidate they want to vote for, and the vote Fig. 1. Block Diagram of Proposed System
is recorded electronically. Once the voting process is
complete, the votes will be counted under the supervision of MODULES DESCRIPTION
ICE officials. The results will then be announced for each
constituency. 1. USER MODULE:

The user module within the envisioned E-voting

III. PROPOSED SYSTEM ARCHITECUTRE system utilizing face recognition technology plays a pivotal
An E-voting System using Face Recognition and OTP role in ensuring a smooth, secure, and user-friendly electoral
verification using HAAR Cascade Classifiers is an innovative
experience. This module facilitates user registration, linchpin of user authentication, enabling precise and real-
enabling individuals to create accounts by providing time verification based on facial attributes. Operating
personal details and capturing facial images for biometric seamlessly with the trained model, this module swiftly
authentication. Subsequent logins involve real-time facial processes facial images captured during login or voting. By
recognition for robust user verification. Once authenticated, extracting distinctive features and generating unique facial
users interact with the system to cast their votes, access representations, the module establishes a basis for
voter information, and receive timely notifications comparison against registered users' data. Employing a
throughout the process. The module also encompasses user predefined threshold for similarity, the module makes swift
support, data security measures, and avenues for user and accurate authentication determinations. Robust error
feedback, all contributing to a comprehensive and handling mechanisms further refine the recognition process,
trustworthy platform. By prioritizing user convenience and accounting for variations in lighting, poses, and expressions.
security, this module underscores the project's commitment Through these mechanisms, the module not only assures
to modernizing elections and fostering a sense of security and accuracy but also ensures an enhanced user
engagement, trust, and empowerment among voters. experience, enabling efficient and seamless participation in
the electoral process.
2. ADMIN MODULE:
5. OTP GENERATION:
The admin module within the envisioned E-voting
system, enriched by face recognition technology, embodies The OTP era module inside the imagined E-
the backbone of system administration, overseeing critical voting framework utilizing confront acknowledgment
components for a seamless and secure electoral process. innovation may be an imperative component that improves
This module empowers administrators to effectively manage security through multi-factor verification. When clients start
user accounts, supervise candidate registrations, and control login or basic activities, this module produces special, time-
the voting process timeline. By ensuring security through sensitive OTPs based on client personality and a mystery
activity monitoring, audit logs, and data management, key. Transmitted safely to the client, the OTP serves as a
administrators maintain the system's integrity. The module one-time get to code. Clients input this code nearby their
also facilitates the compilation and verification of election real-time captured facial picture, supporting verification. By
results, driving transparency and credibility. Through real- approving the OTP and facial acknowledgment, the module
time monitoring, communication with users, and system gifts secure get to. Time-limited OTP legitimacy and anti-
maintenance, administrators guarantee the system's stability fraud measures assist invigorate the system. This module's
and performance. Ultimately, the admin module acts as a integration strengthens believe within the system's security
pivotal force in orchestrating a modern, efficient, and and adjusts with the project's commitment to giving a secure
trustworthy E- voting system, aligned with the demands of and reliable E-voting encounter.
advanced technology and democratic ideals.
6. VOTING SYSTEM:
The voting module within the envisioned E-
voting system using face recognition technology serves as
3. MODLE TRAINING:
the pivotal gateway for users to exercise their democratic
rights securely and efficiently. Upon successful
The model training module within the envisioned
authentication, users are presented with a ballot featuring
E- voting system employing face recognition technology
relevant candidates and measures. User-friendly interfaces
serves as the bedrock for precise and reliable user
facilitate candidate selection and offer confirmation
authentication. Starting with the assembly of a diverse
opportunities before securely recording votes. The module
dataset representing registered voters, the module undergoes
ensures vote anonymity, employing encryption and robust
a series of pivotal steps. Data preprocessing optimizes
security measures to safeguard voter data. By maintaining
images for analysis, while advanced feature extraction,
real-time transparency and feedback, users are assured of the
namely LBPH Face recognizer, deciphers intricate facial
accuracy of their selections. The module's audit trail
patterns. The model architecture, comprising numerous
enhances accountability and contributes to the accurate
layers, progressively learns facial features, refining its
compilation of results. Ultimately, this module embodies the
understanding with each iteration. Through iterative training
core essence of the E-voting system, streamlining the voting
and the tuning of hyperparameters, the model refines its
process and fostering a trustworthy, accessible, and
accuracy. By minimizing a defined loss function, the model
participatory electoral experience.
learns to differentiate individuals based on facial attributes.
Rigorous evaluation ensures the model's robustness and PROPOSED SYSTEM ALGORITHM
effectiveness, reinforcing its role as a cornerstone in
securing the E-voting system through accurate and
trustworthy user authentication. 1. HAAR CASCADED CLASSIFIER :
A HAAR Cascaded Classifier may be a machine
4. FACE RECOGNITION: learning protest discovery method utilized in computer
The face recognition module within the envisioned vision. It's planned to effectively and precisely distinguish
E-voting system using face recognition technology is the particular objects or designs inside pictures or video
streams. It works
by breaking down the discovery handle into stages, with
each arrange comprising of a "cascade" of straightforward
classifiers. These classifiers utilize HAAR-like highlights to
recognize between question and non-object districts,
permitting for quick dismissal of non-object locales,
decreasing computational stack. In each organize, the
classifier evaluates a subset of highlights, and on the off
chance that the locale doesn't coordinate the object's
characteristics, it's rapidly disposed of. This cascading
approach increments proficiency, making it appropriate for
real-time applications like confront discovery. HAAR
Cascaded Classifiers have been utilized broadly in
applications like confront location in cameras and facial
acknowledgment frameworks.

2. HAAR CASCADED CLASSIFIER FOR FACE

RECOGNITION:
The HAAR calculation could be a prevalent
strategy for confront location and acknowledgment in
images and recordings, and it can be utilized for numerous
applications. The primary step in is to distinguish the faces
within the pictures . For this, ready to utilize the HAAR
cascade classifier, which may be a pre-trained machine
learning demonstrate that can distinguish faces based on a
set of HAAR-like highlights. The classifier is prepared on a
expansive dataset of positive (faces) and negative (non-
faces) pictures. To utilize the HAAR classifier in Python,
you'll be able utilize the OpenCV library, which gives a pre-
trained HAAR classifier for confront detection. Here's a few
test code to identify faces utilizing the HAAR classifier:
Distinguishing a given protest in an picture is known as
protest discovery. This errand can be accomplished utilizing
a few procedures, but in this article, we'll utilize HAAR
Fig.2. FLOWCHART FOR FACE DETECTION
cascade with pre-trained XML records. It's the best method
to perform question discovery. HAAR cascades have been
utilized for protest location on low-edge gadgets, and it was
one of the foremost prevalent question location calculations IV. WORKING FLOW OF HAAR CASCADE
in OpenCV. HAAR Cascade isn't much computation-heavy; ALGORITHM:
subsequently it is prevalent for little gadgets with little One of the popular algorithms for facial detection is
computation control. HAAR Cascade could be a feature- “haar cascade”. It is computationally less expensive, a fast
based question location algorithm to detect objects from algorithm, and gives high accuracy.
pictures. A cascade work is prepared on parcels of positive
and negative pictures for discovery. The algorithm does not Haar feature selection:
require broad computation and can run in real-time. We will Haar feature selection is a crucial step in the Haar
prepare our own cascade work for custom objects like Cascade Object Detection framework, which is widely used
Human confront, creatures, cars, bicycles, etc. HAAR for detecting objects, particularly faces, in images or video.
Cascade can't be utilized for confront acknowledgment since In Haar feature selection, a set of features is chosen to
it as it were recognizing the coordinating shape and efficiently and accurately identify objects of interest within
measure. Haar cascade employments the cascade work and images.
cascading window. It tries to calculate highlights for each Feature Engineering in Haar Cascade: The Haar Cascade
window and classify positive and negative. In case the Object Detection framework utilizes a machine learning
window can be a portion of an protest, at that point positive, approach to identify objects. Central to this framework is the
else, negative. concept of Haar-like features, which are simple rectangular
patterns used to distinguish between object and non-object
regions in an image.
Haar-Like Features: Haar-like features are defined by their
shape, size, and position within an image. These features
can take on various forms, such as rectangular boxes or
tilted
rectangles. Haar-like features capture information about the Integral images eliminate the need to repeatedly compute
distribution of pixel intensities within their defined region. the sum of pixel values within regions, as these sums are
Integral Image: To efficiently compute Haar-like features, an precomputed and stored. This leads to substantial time
integral image is constructed from the original image. The savings.
integral image allows for rapid summation of pixel values Creating an integral image involves iterating through the
within rectangular regions, making feature computation original image pixel by pixel and applying a cumulative sum
much faster. operation. The process starts at the top-left corner of the
Feature Types Haar-like features come in different types, image and proceeds row by row and column by column,
such as edge features, line features, and center-surround adding up the pixel values as it goes. It can be efficiently
features. These features help capture specific patterns and implemented in a single pass, making it computationally
textures within the objects of interest. inexpensive.
Haar feature selection is part of the training process for a
Haar Cascade classifier. During training, a machine learning CONCULSION
algorithm like Adaboost is employed. Adaboost selects the In this paper presents a promising approach to enhance the
most discriminative Haar-like features from a vast pool of security, efficiency, and integrity of electoral processes. By
potential features, giving higher importance to those that leveraging the uniqueness of facial features, this technology
contribute most to accurate object detection. offers robust voter authentication, minimizing the risk of
In the training process, positive samples (images containing fraud and duplicate voting. The convenience of facial
the object of interest) and negative samples (images without recognition streamlines the voting experience for citizens,
the object) are used. The classifier learns to distinguish while HAAR Cascade Classifiers ensure accurate and
between these two classes based on the selected Haar-like reliable facial pattern analysis. However, successful
features. implementation requires careful consideration of privacy
After feature selection and training, the classifier is protection, addressing biases, and maintaining transparency
organized into a cascade structure, consisting of multiple to gain public trust. With appropriate measures, e-voting
stages, each with several weak classifiers. These stages are with Face Recognition and OTP verification can pave the
arranged in a way that quickly eliminates non-object way for more accessible and secure democratic elections in
regions, optimizing detection speed. the digital era.
For each Haar-like feature, a threshold value is determined
during training. The classifier compares the feature's REFERENCES
computed value to this threshold to decide whether a region [1] U. Can Cabuk, E. Adiguzel, and E. Karaarslan, ‘‘A survey on
is likely to contain the object. feasibility and suitability of blockchain techniques for the E-voting
Once trained, the Haar Cascade classifier can be applied to systems,’’ 2020, arXiv:2002.07175.
new images or video frames to detect the presence of [2] J. Ben-Nun, N. Fahri, M. Llewellyn, B. Riva, A. Rosen, A. Ta-
objects. The cascade structure, along with feature selection Shma, and D. Wikström, ‘‘A new implementation of a dual (paper
and thresholding, allows for real-time or near-real-time and cryptographic) voting system,’’ in Proc. 5th Int. Conf.
object detection with high accuracy. Electron. Voting (EVOTE),2012, pp. 315–329.
[3] S. K. Vivek, R. S. Yashank, Y. Prashanth, N. Yashas, and M.
Namratha,‘‘E-voting systems using blockchain: An exploratory
Creation of Integral Images: literature survey,’’ in Proc. 2nd Int. Conf. Inventive Res. Comput.
An integral image, also known as a summed-area Appl. (ICIRCA), Jul. 2020,pp. 890–895 Taş and Ö. Ö. Tanriöver,
table, is a representation of an original image where each ‘‘A systematic review of challenges and opportunities of
blockchain for E-voting,’’ Symmetry, vol. 12, no. 8,p. 1328, Aug.
pixel contains the sum of all pixel values located above and 2020.
to the left of it. In mathematical terms, if we have an image [4]S. Heiberg, K. Krips, J. Willemson, and P. Vinkel, ‘‘Facial
with pixel values represented as I (x, y), the integral image recognition for remote electronic voting–missing piece of the
at position (x, y) is computed as: puzzle or yet another liability?’’ in Proc. Int. Workshop Emerg.
Integral (x, y) = ∑∑ I (i, j) for i ≤ x and j ≤ y Technol. Authorization Authentication. Switzerland: Springer,
This process essentially creates a cumulative distribution of 2021, pp. 77– 93
pixel values, which can be efficiently queried to calculate [5] U. Jafar, M. J. A. Aziz, and Z. Shukur, ‘‘Blockchain for
the sum of pixel values within any rectangular region in electronic voting system—Review and open research challenges,’’
Sensors, vol. 21, no. 17, p. 5874, 2021.
constant time, regardless of the region's size.
[6] S. Al-Maaitah, M. Qatawneh, and A. Quzmar, ‘‘E-voting
Integral images make it extremely efficient to compute the system based on blockchain technology: A survey,’’ in Proc. Int.
sum of pixel values within rectangular regions, which is a Conf. Inf. Technol. (ICIT),Jul. 2021, pp. 200–205.
common operation in various image processing tasks. [7] Y. Abuidris, R. Kumar, and W. Wenyong, ‘‘A survey of
Integral images are particularly valuable in speeding up blockchain based on E-voting systems,’’ in Proc. 2nd Int. Conf.
feature extraction for object detection algorithms like Haar Blockchain Technol. Appl., Dec. 2019, pp. 99–104
cascades and Viola-Jones. These algorithms use rectangular [8]K. T. Sri, K. R. Sri, and N. Pedamallu, ‘‘E-voting system using
filters that slide over an image, and integral images help blockchain,’’ J. Xi’an Univ. Archit. Technol., vol. 13, no. 5, pp.
compute the filter responses efficiently. 527–533, 2021.
[9] V. Anilkumar, J. A. Joji, A. Afzal, and R. Sheik, ‘‘Blockchain
For real-time applications such as face detection in video
simulation and development platforms: Survey, issues and
streams, integral images are crucial. They significantly challenges,’’in Proc. Int. Conf. Intell. Comput. Control Syst.
reduce the computational burden, enabling faster and more (ICCS), May 2019,pp. 935–939.
responsive systems. [10] T. Moura and A. Gomes, ‘‘Blockchain voting and its effects
on election transparency and voter confidence,’’ in Proc. 18th
Annu. Int. Conf. Digit.Government Res., Jun. 2017, pp. 574–575.