SRM INSTITUTE OF SCIENCE AND TECHNOLOGY

FACULTY OF ENGINEERING AND TECHNOLOGY

DEPARTMENT OF NETWORKING AND COMMUNICATIONS
18CSP109L- MAJOR PROJECT

Streamlined crime reporting and an AI integrated cloud

based system for enhanced community safety

Student 1 Reg No: RA2011028010017

Student 2 Reg No: RA2011028010019
Batch ID:
Panel No:
Guide name and Designation: Dr.
 Table of Contents
● Abstract
● Introduction
● Objectives
● Literature Survey
● Proposed Architecture
● Proposed Modules
● Conclusion
● References
● Plan Of Action
 Abstract

The project introduces an innovative online system leveraging AI integration within a

cloud-based framework for streamlined crime reporting. With a primary focus on
enhancing community safety, this platform empowers citizens to efficiently report
crimes, fostering collaboration between the public and law enforcement. Through the
application of advanced technology, the system facilitates swift data analysis and
response, contributing significantly to the overall improvement of community safety.
 Introduction
• Our project responds to the shifting landscape of public safety by introducing an
innovative online system for streamlined crime reporting. Leveraging artificial
intelligence (AI) within a cloud-based infrastructure, our goal is to simplify and
enhance the crime reporting process.

• Our user-friendly platform empowers citizens to seamlessly report crimes, while AI

technologies enable intelligent data analysis and immediate responses.Operating on
the cloud, our system transcends geographical limitations, fostering collaboration
between citizens and law enforcement.

• This introduction emphasizes our project's motivation, highlighting the demand for
proactive and technologically advanced solutions to strengthen community safety.
 Literature Survey
• This paper in [1] systematically reviews the use of machine learning and deep
learning in predicting crimes, examining existing research, methodologies, and
outcomes. It also forecasts future directions, offering insights into potential
advancements and unexplored territories in crime prediction through advanced
computational approaches.

• This paper in [2] assesses i-recall, an online crime reporting system, by contrasting
it with a non-interactive reporting method. i-recall imitates a police officer's
cognitive interview, incorporating cognitive techniques and natural language
processing to improve witness memory recall. The study involved college students
witnessing a staged crime and reporting their observations via i-recall or a non-
interactive textbox. Results show i-recall's superiority in completeness and accuracy,
indicating its potential as an effective internet crime reporting tool.
• This paper proposed in [3] introduces "e-Cops," an online system for reporting and
managing crimes in Riyadh. It is designed as a digital platform to improve the
efficiency of crime reporting in the city. It probably delineates e-Cops' features,
focusing on its role in enabling online crime reporting, streamlining management,
and potentially enhancing law enforcement-public communication. The aim is to
enhance the city's crime management system by leveraging online technology

• This paper proposed in [4] presents an automated system for reporting crimes and
initiating immediate responses. The system aims to automate crime reporting and
ensure swift responses. The primary focus is on utilizing technology to improve
the efficiency of reporting crimes, enabling rapid and automated communication
with law enforcement. The objective is to create a system that minimizes reporting
delays and enables swift responses to reported incidents, enhancing public safety
and law enforcement effectiveness.
 Objectives
• Develop an Efficient Crime Reporting Mechanism: The primary objective is to create a
streamlined and efficient mechanism for reporting crimes automatically, reducing delays in
the reporting process

• Utilize Automation Technology: Incorporate cutting-edge automation technology to enable

seamless and rapid crime reporting, reducing reliance on manual reporting methods and
improving overall system efficiency.

• Enhance Communication with Law Enforcement: Improve communication channels

between the public and law enforcement by implementing an automated crime reporting
system, fostering quicker and more effective responses to reported incidents.
• Ensure User-Friendly Interface: Design a user-friendly interface for the crime
reporting system to encourage widespread adoption and ease of use by the public.

• Establish Protocols for Immediate Action: Develop protocols and procedures to ensure
that reported crimes trigger immediate and appropriate responses, contributing to the
prevention and resolution of incidents.

• Incorporate Security Measures: Implement robust security measures within the system
to safeguard sensitive information, ensuring the confidentiality and integrity of reported
crime data.
 Proposed Architecture
• User Interface Layer: Retain HTML for structure and Bootstrap for responsive
design.Utilize JavaScript for enhanced user interactivity.

• Application Layer: Maintain Python as the primary language for server-side logic.
Continue using Flask as the web framework to handle user requests and manage
interactions with the database.

• Database Layer: Retain SQLite for its simplicity and portability

• Analysis and Visualization Layer: Continue using Pandas and Matplotlib for data
analysis and visualization.
• Machine Learning Layer: Integrate Natural Language Processing (NLP) and Scikit-
learn for advanced analysis and prediction of criminal behaviour based on reported
incidents.

• Facial Recognition Layer: Integrate OpenCV and LBPH Face Recognizer for facial
recognition to identify individuals involved in criminal activities.

• Additional Considerations: Enhance scalability by considering a more robust relational

database system (e.g., PostgreSQL or MySQL) if scalability becomes a concern.

• Implement security measures, including encryption and secure communication

protocols, to safeguard sensitive information in the system.
System Architecture:
• User Interface (UI): Components: HTML, Bootstrap, JavaScript.
Description: The front-end UI provides a user-friendly interface for citizens to report
crimes seamlessly. It includes forms, interactive elements, and design elements
facilitated by HTML, Bootstrap for responsive layouts, and JavaScript for
interactivity.

• Backend Services:Framework: Python with Flask.

Functionality: Flask-based backend services handle user requests, process data, and
manage the core functionalities of the crime reporting system. Python scripts manage
routing, handle database interactions, and process user inputs.

• Database Management:Database: SQLite.

Role: SQLite serves as the database management system, storing reported crime data
securely. It manages data integrity, storage, and retrieval for efficient access during
crime investigations.
• Data Analysis and Visualization:
Libraries: Pandas, Matplotlib.
Purpose: Pandas supports data manipulation and analysis, while Matplotlib aids in
generating visual representations of crime data. These tools help in deriving insights
from reported incidents.

• Machine Learning Integration:Libraries: Scikit-learn, NLP Techniques.

Functionality: Scikit-learn facilitates machine learning tasks, such as crime prediction
or classification. NLP techniques assist in text analysis within crime reports, enabling
sentiment analysis or keyword extraction for improved data understanding.

• Face Detection and Recognition:

Tools: OpenCV, LBPH Face Recognizer.
Usage: OpenCV powers face detection and recognition capabilities within the system.
It aids in identifying suspects or verifying identities based on images or video footage.
• Integration Layer:
Purpose: The integration layer facilitates communication between various system
components, ensuring seamless data exchange between the UI, backend services, database,
analysis modules, and machine learning models.

Security Measures:
Implementation: Robust security measures, including encryption, access controls, and secure
authentication, are embedded across the architecture to protect sensitive crime-related data
and ensure data privacy.
USE CASE DIAGRAM
Architectural Diagram
 Algorithms To be Used
Logistic Regression: Predicts the likelihood of a binary outcome, such as whether a crime will occur
in a specific location based on various features.

Random Forest: Utilizes multiple decision trees to predict crime by considering complex
interactions between features.

Support Vector Machines (SVM): Separates different crime types based on distinctive features in a
higher-dimensional space.

K-Nearest Neighbors (KNN): Identifies areas with similar crime patterns by measuring the
proximity of instances in a feature space.

Neural Networks: Deep learning models capable of handling complex patterns in crime-related data,
such as images, text, or sequential data.
 Algorithms To be Used
Decision Trees: Hierarchical models that identify important factors contributing to criminal
activities, aiding in interpretation.

Clustering Algorithms (e.g., K-Means): Identifies spatial clusters in crime data, revealing hotspots
or areas with similar crime characteristics.

Ensemble Methods: Combines multiple models for enhanced predictive accuracy and model
stability in crime prediction tasks.
OpenCV: OpenCV provides pre-trained models and libraries for face detection, and it may involve
algorithms like Haar cascades or deep learning-based methods for facial recognition.

LBPH Face Recognizer: Local Binary Pattern Histograms (LBPH) is a feature-based face
recognition algorithm often used in conjunction with OpenCV for facial recognition tasks.
 Evaluation Metrics
Evaluation Metrics

• Accuracy:Accuracy measures the overall correctness of predictions made

by a system. It calculates the proportion of correctly classified instances
among all instances.

• Precision:Precision focuses on the accuracy of positive predictions,

indicating the ratio of correctly predicted positive observations to the
total predicted positives.

• Recall (Sensitivity):Recall assesses the system's ability to identify all

relevant instances by measuring the ratio of correctly predicted positive
observations to all actual positives.
• F1 Score:
Description: F1 Score is the harmonic mean of precision and recall. It provides a
balanced assessment, especially in scenarios where both precision and recall are
essential.

• Mean Absolute Error (MAE) and Mean Squared Error (MSE):

Description: MAE measures the average absolute difference between predicted
and actual values, while MSE measures the average squared difference.
 Proposed Modules

1.User Interface Module: This module constitutes the frontend interface

accessible to users for reporting crimes. It offers a user-friendly platform
enabling citizens to submit crime reports seamlessly.
2.Crime Data Management Module: Centralizes the storage and management
of reported crime data. This module ensures secure handling, organization,
and retrieval of incident records.
3.Analytics and Prediction Module: Utilizes machine learning and analytics to
predict crime patterns, trends, and hotspots based on historical data analysis.
4. Administration and Access Control Module:
Description: Governs system access, user roles, and administrative
functionalities to ensure secure and authorized usage of the system.

5. Geographic Information System (GIS) Integration Module:

Description: Integrates GIS functionalities to map reported crimes,
visualize crime patterns, and geographically analyze crime data.

6. Reporting and Documentation Module: Manages the generation,

storage, and retrieval of comprehensive crime reports, aiding in case
documentation and legal proceedings.
 Conclusion
The "City Without Crime" project aims to create a safer society by leveraging
technology to streamline crime reporting, analysis, and prediction. The use of
various technologies empowers law enforcement and engages the public in a
collaborative effort towards a crime-free community. Continuous improvement
and innovation are essential for achieving the project's mission and ensuring a
safer future for all citizens.
 References
• N. Shah, N. Bhagat and M. Shah, "Crime forecasting: A machine learning and
computer vision approach to crime prediction and prevention", Vis. Comput. Ind.
Biomed. Art, vol. 4, no. 1, pp. 1-14, Apr. 2021.

• S. A. Chun, V. A. Paturu, S. Yuan, R. Pathak, V. Atluri and N. R. Adam, "Crime

prediction model using deep neural networks", Proc. 20th Annu. Int. Conf. Digit.
Government Res., pp. 512-514, Jun. 2019.

• S. S. Kshatri, D. Singh, B. Narain, S. Bhatia, M. T. Quasim and G. R. Sinha, "An

empirical analysis of machine learning algorithms for crime prediction using
stacked generalization: An ensemble approach", IEEE Access, vol. 9, pp. 67488-
67500, 2021.
• W. Safat, S. Asghar and S. A. Gillani, "Empirical analysis for crime prediction
and forecasting using machine learning and deep learning techniques", IEEE
Access, vol. 9, pp. 70080-70094, 2021.

• S. Kim, P. Joshi, P. S. Kalsi and P. Taheri, "Crime analysis through machine

learning", Proc. IEEE 9th Annu. Inf. Technol. Electron. Mobile Commun.
Conf. (IEMCON), pp. 415-420, Nov. 2018.

• S. Hossain, A. Abtahee, I. Kashem, M. M. Hoque and I. H. Sarker, "Crime

prediction using spatio-temporal data", Computing Science Communication
and Security, pp. 277-289, 2020.
• A. Kumar, A. Verma, G. Shinde, Y. Sukhdeve and N. Lal, "Crime prediction
using K-nearest neighboring algorithm", Proc. Int. Conf. Emerg. Trends Inf.
Technol. Eng., pp. 1-4, Feb. 2020.

• X. Han, X. Hu, H. Wu, B. Shen and J. Wu, "Risk prediction of theft crimes in
urban communities: An integrated model of LSTM and ST-GCN", IEEE
Access, vol. 8, pp. 217222-217230, 2020.
 Plan of
Review Number Action
Work to be completed Percentage Dates (Tentative)

Project Review - 0 Abstract,PPT, Literature 5% 20-11-2023

Survey

Project Review - I Detailed Literature 30% 01-01-2024

survey, Initial Report,
Implementation (30%)

Project Review - II Implementation (70%), 70% 01-02-2024

Rough draft - final
report, Research paper

Project Review - III Implementation 100% 01-03-2024

(100%), Final Report,
Final Research Paper,
Publication
Thank
s