Volume 9, Issue 4, April – 2024 International Journal of Innovative Science and Research Technology

ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/IJISRT24APR1653

Criminal Investigation Tracker with

Suspect Prediction
T Sam Pradeep Badugula Tejavardhan Reddy
Department of Computer Science and Engineering Department of Computer Science and Engineering
Kalasalingam Academy of Research and Institute Kalasalingam Academy of Research and Institute
Anand Nagar, Krishnankoil, Tamilnadu Anand Nagar, Krishnankoil, Tamilnadu

Jagarapu Manikanta Voleti Dinesh Kumar

Department of Computer Science and Engineering Department of Computer Science and Engineering
Kalasalingam Academy of Research and Institute Kalasalingam Academy of Research and Institute
Anand Nagar, Krishnankoil, Tamilnadu Anand Nagar, Krishnankoil, Tamilnadu

Sannapaneni SaiTheja
Department of Computer Science and Engineering
Kalasalingam Academy of Research and Institute
Anand Nagar, Krishnankoil, Tamilnadu

Abstract:- This initiative seeks to craft an inclusive and I. INTRODUCTION

intuitive web-based platform tailored for criminal
identification via image and video surveillance, leveraging This study introduces a desktop application that
cutting-edge facial recognition technology. The system surpasses current methodologies in swiftly and accurately
integrates a registration portal for inputting data and identifying offenders based on facial recognition, fingerprint
images of known criminals, utilizing OpenCV and analysis, and weapon detection. Presently, Sri Lanka lacks a
advanced facial recognition algorithms to securely comprehensive system harnessing advanced technology to
analyze and store their facial attributes. In addition to promptly identify illegal activities and alert authorities. The
allowing users to upload images for analysis, the system proposed system will gather and archive crime data, taking
offers immediate feedback on potential matches with contextual factors into consideration to inform relevant parties
registered criminals. Moreover, the video surveillance and the community about incidents in real-time. Additionally,
module extends this capability to short videos, employing it will analyze and append information for future reference. A
video analytics to identify faces within the footage. The team is tasked with devising and implementing a sophisticated
platform ensures real-time feedback for successful approach to compile and analyze data, ensuring sound
identifications and provides an advanced feature enabling conclusions in this expansive field of research. The Sri
users to download details of identified criminals in an Lankan Penal Code was initially enacted to address
Excel format. By amalgamating state-of-the-art deficiencies in the nation's pre-existing criminal legislation,
technology with an intuitive interface, this project and despite updates, remains in force today. Legal defenses
endeavors to bolster law enforcement endeavors by such as insanity, intoxication, necessity, coercion, and self-
furnishing an efficient and precise tool for criminal defense are recognized within this framework. The concept of
identification and tracking. Its objectives encompass 'mens rea,' or criminal intent, is deemed absent in children
developing a robust system for identifying and tracking under the age of eight. Crime, fundamentally, is an act that
criminals through advanced facial recognition algorithms contravenes a law prohibiting it or fails to comply with a law
and OpenCV technology, designing a user-friendly web mandating it. Perpetrators of crime may face repercussions
interface for seamless navigation across various modules, from law enforcement and the government, given its unlawful
and establishing a secure and efficient registration section and reprehensible nature. A criminal is someone who has
for compiling comprehensive databases of facial features. committed or is in the process of committing a crime. Crime
manifests as a societal scourge with multifaceted negative
Keywords:- OpenCV, Facial Identification, State-of-the-Art, impacts. Criminality is increasingly pervasive within society,
User-Friendly Interface. with a small fraction of offenders responsible for a significant
proportion of offenses. Crime is inherently unpredictable,
transcending spatial and temporal boundaries, contingent
upon individual actions. Criminal activities span a spectrum
from minor traffic violations to terrorist acts, generating
copious amounts of data in various forms. Consequently,
evaluating crime data can be challenging, prompting the

IJISRT24APR1653 www.ijisrt.com 1621

Volume 9, Issue 4, April – 2024 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/IJISRT24APR1653

adoption of data mining techniques to derive valuable insights networks, and clustering algorithms, and their applicability in
from vast datasets. Despite advancements in technology, law crime prediction, offering insights into their strengths and
enforcement's utilization of such tools for crime prevention limitations across different crime types.
and offender identification lags behind the sophistication of
criminal tactics. In response to the inherent challenges of B. "Evaluating the Efficacy of Predictive Policing: A
criminal investigations, we present a solution: the Criminal Systematic Review"
Investigation Tracker with Suspect Prediction. This
innovative system leverages machine learning algorithms to Journal of Police and Criminal Psychology
redefine conventional investigative methods, empowering law M. J. Gerber, T. M. Dowling
enforcement agencies with unparalleled efficiency and 2020
accuracy. By harnessing predictive modeling, data analytics, This systematic review scrutinizes empirical studies on
and advanced algorithms, our solution facilitates proactive predictive policing, assessing the effectiveness of predictive
identification of potential suspects and actionable intelligence analytics in crime prevention and law enforcement, while
throughout the investigative process. The core functionality of also addressing methodological nuances and ethical
our solution revolves around predictive modeling, which considerations.
analyzes historical crime data and behavioral patterns to
forecast potential suspects based on predefined criteria. C. "Exploring Predictive Policing Approaches: A Survey"
Through continuous surveillance and analysis of incoming
data streams, our system detects suspicious activities in real- International Journal of Law, Crime and Justice
time, enabling prompt intervention. Moreover, it integrates S. K. Chauhan, A. K. Kadian
diverse data sources and presents actionable insights through 2020
intuitive dashboards, fostering collaboration and information Through a survey approach, this paper investigates
sharing among law enforcement agencies. Key features of our diverse predictive policing methodologies, encompassing
solution include: hotspot analysis, risk terrain modeling, and machine learning
 Predictive Suspect Identification: Utilizing machine algorithms, offering a critical evaluation of their merits and
learning to identify individuals with a higher likelihood of demerits, alongside discussions on ethical and privacy
involvement in criminal activities. concerns.
 Real-time Monitoring: Continuous surveillance and
analysis of data streams to detect emerging patterns and D. "Assessing Crime Prediction Models: A Comprehensive
anomalies. Review"
 Data Integration and Visualization: Aggregating diverse
data sources and presenting insights through intuitive Decision Support Systems
dashboards. A. K. Jha, R. F. Ahmed
 Collaboration Tools: Facilitating communication and 2020
information sharing among law enforcement personnel to This review paper evaluates existing crime prediction
enhance coordination. models and techniques, spanning traditional statistical
 Scalability and Flexibility: Adaptable architecture capable methods and modern machine learning algorithms, while
of handling large volumes of data and users, ensuring delineating methodological frameworks and addressing
scalability and flexibility in operations. challenges such as data quality issues and model
interpretability.
II. LITERATURE SURVEY
E. "Navigating the Landscape of Predictive Policing: A
The literature review highlights similar themes, delving Review and Outlook"
into research roles and components pertinent to the subject
matter. The primary aim of this study is to develop an European Journal of Criminology
entirely automated criminal investigation and tracking A. M. Posada-Quintero, L. W. Kennedy
system, incorporating a suspect prediction mechanism 2019
utilizing CCTV imagery from public areas. This core system Offering a panoramic view, this paper surveys the
consists of four integral components: the face detection literature on predictive policing, examining future trajectories
system, weapon detection and alarm triggering systems, for research and implementation, while addressing
fingerprint detection system, and smart notification system, methodological hurdles such as data biases, algorithmic
alongside the content summarizer. transparency, and community involvement.

A. "Exploring Data Mining Techniques for Crime F. "Unveiling Predictive Analytics and Crime Mapping in
Prediction: A Review" Law Enforcement: A Comprehensive Review"

Expert Systems with Applications Policing: An International Journal

S. F. Al-Shatnawi, A. M. Al-Sayyed M. Haberman, K. Ratcliffe
2020 2021
This comprehensive review paper examines various This paper scrutinizes law enforcement practices and
data mining methodologies, including decision trees, neural research concerning predictive analytics and crime mapping,

IJISRT24APR1653 www.ijisrt.com 1622

Volume 9, Issue 4, April – 2024 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/IJISRT24APR1653

discussing methodological paradigms in predictive modeling  Modules

and identifying challenges like reliance on historical data and
predictive accuracy assessment issues.  User Interface:
 Craft an intuitive web interface to facilitate seamless user
III. METHODOLOGY interaction.
 Incorporate dedicated sections for registration, image
A. "Exploration of Crime Prediction through Data Mining surveillance, and video surveillance.
Techniques: A Systematic Literature Review"  Ensure real-time visual feedback upon successful
 Methodology: Systematically conducted a review of identifications.
literature focusing on crime prediction utilizing data  Implement functionality to securely store registered
mining techniques. Evaluated a broad spectrum of studies criminals' details in an Excel file.
to assess the efficacy of different data mining algorithms
in predicting various types of crimes.  Face Recognition Module:
 Harness the power of OpenCV and advanced facial
B. "Evaluation of Predictive Policing: A Systematic Review recognition algorithms for image analysis.
of Empirical Evidence"  Capture and securely store facial features of registered
 Methodology: Systematically reviewed empirical studies criminals for reference.
concerning predictive policing to gauge the substantiation  Develop an efficient matching system to ascertain the
supporting its efficacy. Assessed methodological presence of known criminal faces in uploaded images.
approaches employed in these studies, encompassing
experimental designs, data collection methodologies, and  Image Surveillance Module:
statistical analysis techniques.
 Enable users to submit images for thorough analysis.
 Utilize the Face Recognition Module to swiftly identify
C. "Surveying Predictive Policing Approaches for Crime
potential matches with registered criminals.
Analysis"
 Methodology: Conducted a survey encompassing
 Advanced Video Surveillance Module:
predictive policing methodologies, including hotspot
 Empower users to upload brief videos for comprehensive
analysis, risk terrain modeling, and machine learning
analysis.
algorithms. Evaluated the strengths and weaknesses of
each approach based on existing literature, discussing  Deploy sophisticated video analytics to detect and extract
challenges such as algorithmic fairness and privacy facial features from the footage.
concerns.  Employ the Face Recognition Module to identify any
registered criminals captured within the video content.
D. "Assessment of Crime Prediction Models: A criminals.
Comprehensive Review"
 Methodology: Reviewed existing crime prediction models IV. RESULTS AND DISCUSSION
and techniques, spanning statistical methods and machine
learning algorithms. Examined methodological approaches The system conducts a comparison between the
utilized in the development of these models, including provided image and those within the database, evaluating the
data preprocessing, feature selection, and model similarity rate. If the matching rate exceeds 80%, the
evaluation. algorithm identifies the individual as a suspect, prompting
the display of relevant information such as name, age, and ID
E. "Analysis of Predictive Policing: Review and Future card number. In cases where no match is found, the system
Perspectives" archives the supplied image as a new entry. Notably, the
 Methodology: Reviewed literature on predictive policing system's advancements enable it to discern profiles even from
and deliberated on future research directions. Examined side-facing images and to enhance recognition capabilities in
methodological hurdles in predictive policing research, videos with poor visibility.
such as data quality concerns, algorithmic transparency,
and community engagement. On the other hand, summarizing police reports or
records presents a significant efficiency challenge.
F. "Reviewing Predictive Analytics and Crime Mapping in Traditional reports often contain lengthy paragraphs and
Law Enforcement Practices and Research" intricate details, demanding considerable time for
Methodology: Reviewed law enforcement practices and comprehension. To address this, our system employs
research concerning predictive analytics and crime mapping. automated summarization techniques. Upon inputting a
Examined methodological approaches in predictive modeling, police report, the system condenses the content, extracting
including data preprocessing, spatial analysis, and model essential information and presenting it to officers in a concise
evaluation. Identified limitations and challenges in format. This streamlined approach significantly reduces the
implementing predictive analytics in law enforcement. time required for information processing. Moreover, our
study introduces a novel method utilizing non-fuzzy concepts
to extract critical data from documents, offering a unique

IJISRT24APR1653 www.ijisrt.com 1623

Volume 9, Issue 4, April – 2024 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/IJISRT24APR1653

solution that may evolve to complement emerging intelligent

systems capable of sophisticated textual analysis.

Fig 2: Face Recognition

C. "Exploring Predictive Policing Approaches: Findings

from a Survey"
Fig 1: Register criminal details  Results:
 Recognized hotspot analysis, risk terrain modeling, and
A. "Crime Prediction Using Data Mining Techniques:
machine learning as key predictive policing
Insights from a Literature Review"
methodologies.
 Results:
 Discussion:
 Identified a range of data mining techniques utilized in
 Highlighted the simplicity of hotspot analysis but noted
crime prediction.
its potential limitations in capturing underlying spatial
 Evaluated the efficacy of decision trees, neural networks, patterns effectively.
and clustering algorithms.
 Discussed the incorporation of environmental factors in
risk terrain modeling, alongside the necessity for
 Discussion:
extensive data.
 Decision trees offer interpretability but may sacrifice
 Explored the flexibility of machine learning algorithms
predictive accuracy.
while acknowledging challenges such as algorithmic bias
 Neural networks demonstrate high predictive accuracy and the risk of overfitting.
but are often perceived as opaque "black boxes."
 Clustering algorithms aid in identifying crime patterns but
may not excel in predictive modeling.

B. "Predictive Policing: Analyzing Empirical Evidence

through a Systematic Review"

 Results:
 Examined empirical evidence regarding the effectiveness
of predictive policing strategies.
 Investigated the methodological approaches employed
across the reviewed studies.

 Discussion:
 Addressed the variability in results stemming from
disparities in data quality and model accuracy.
Fig 3: Detect the Criminal Face
 Explored ethical considerations surrounding predictive
policing, including concerns regarding bias and privacy
D. "Examining Crime Prediction Models: Insights from a
implications.
Comprehensive Review"

 Results:
 Evaluated various crime prediction models,
encompassing statistical methodologies and machine
learning approaches.

IJISRT24APR1653 www.ijisrt.com 1624

Volume 9, Issue 4, April – 2024 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/IJISRT24APR1653

VI. FUTURE ENHANCEMENT

 Discussion:
 Noted the interpretability of statistical methods but Looking ahead, there exist numerous avenues for
highlighted potential limitations in predictive accuracy. bolstering the effectiveness and relevance of predictive
 Explored the enhanced predictive accuracy offered by policing and crime prediction models. One promising avenue
machine learning models, alongside the challenges of involves the integration of advanced machine learning
extensive data preprocessing and tuning. methodologies, such as deep learning and reinforcement
 Addressed concerns regarding model interpretability and learning, aimed at refining the accuracy and resilience of
generalizability, emphasizing the need for further predictive models. These methodologies hold the potential to
research in these areas. unveil intricate patterns within crime data and dynamically
refine predictive algorithms over time. Furthermore, the
E. "Assessing Predictive Policing: Insights and Future incorporation of real-time data sources, including social media
Perspectives" feeds, IoT devices, and satellite imagery, could amplify the
promptness and granularity of predictive insights, enabling
 Results: law enforcement agencies to swiftly address emerging threats
 Investigated the existing literature surrounding predictive and evolving crime trends.
policing strategies.
 Identified gaps in both research and practical In essence, forthcoming enhancements in predictive
implementation. policing and crime prediction models should prioritize the
integration of sophisticated machine learning methodologies,
 Discussion: the mitigation of algorithmic bias and fairness concerns, and
 Advocated for more rigorous evaluation studies to the assimilation of predictive analytics into comprehensive
accurately assess the effectiveness of predictive policing. crime prevention strategies. By embracing these strategies and
 Stressed the significance of algorithmic transparency and fostering collaboration across various disciplines and sectors,
community involvement in shaping future research and we can harness the full potential of predictive analytics to
practical applications of predictive policing. cultivate safer, fairer, and more resilient communities.

V. CONCLUSION REFERENCES

In summary, the literature review offers invaluable [1]. Mohler, G. O., Short, M. B., Malinowski, S., Johnson,
insights into the utilization of data mining techniques and M., Tita, G. E., Bertozzi, A. L., & Brantingham, P. J.
predictive policing methodologies in crime prediction and (2015). Randomized controlled field trials of predictive
law enforcement. While methods like decision trees, neural policing. Journal of the American Statistical
networks, and clustering algorithms each bring unique Association, 110(512), 1399-1411.
advantages such as interpretability, accuracy, and pattern [2]. Ratcliffe, J. H. (2016). Intelligence-led policing.
recognition, they also present notable drawbacks like Routledge.
algorithmic bias, data integrity issues, and ethical dilemmas. [3]. Mohler, G. O., & Short, M. B. (2019). Criminal
Furthermore, the empirical evidence underscores the forecasting: past, present, and future. Annals of the
necessity for more stringent evaluation studies and increased American Association of Geographers, 109(2), 446-454.
transparency in algorithmic decision-making to mitigate [4]. Lum, K., & Isaac, W. (2016). To predict and serve?.
concerns surrounding fairness and accountability. Significance, 13(5), 14-19.
[5]. Groff, E. R., & La Vigne, N. G. (2002). Forecasting the
Looking ahead, it is imperative for future research and future of predictive crime mapping. Crime Prevention
practical implementations to prioritize the development of Studies, 13, 71-98.
robust crime prediction models that strike a balance between [6]. Brantingham, P. J., & Brantingham, P. L. (1993).
accuracy and interpretability, while simultaneously Environment, routine, and situation: Toward a pattern
addressing ethical and privacy considerations. Moreover, theory of crime. Advances in criminological theory, 5,
fostering enhanced collaboration among researchers, law 259-294.
enforcement entities, and communities is essential to ensure [7]. Groff, E. R. (2015). Mapping disorder: Theoretical and
that predictive policing initiatives are founded upon empirical challenges. In Environmental Criminology
evidence-based strategies and uphold principles of fairness, and Crime Analysis (pp. 49-76). Routledge.
transparency, and accountability. [8]. Sengupta, S., & Chen, H. (2012). Text mining:
concepts, process and applications. Journal of
By tackling these challenges head-on and harnessing the Computing and Information Technology, 20(1), 55-71.
capabilities of predictive analytics, we can propel [9]. Piquero, A. R., Farrington, D. P., & Blumstein, A.
advancements in law enforcement practices and contribute to (2007). Key issues in criminal career research: New
the establishment of safer and more just communities for all. analyses of the Cambridge Study in Delinquent
Development. Cambridge University Press.
[10]. Ashby, M. P. (2013). Initial conditions and moment
restrictions in dynamic panel data models. Journal of
Econometrics, 175(1), 1-21.

IJISRT24APR1653 www.ijisrt.com 1625

Volume 9, Issue 4, April – 2024 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/IJISRT24APR1653

[11]. Haberman, C. P., & Ratcliffe, J. H. (2015). Testing for

temporally differentiated relationships among
potentially criminogenic places and census block street
robbery counts. Journal of Quantitative Criminology,
31(3), 415-435.
[12]. Anselin, L., Cohen, J., Cook, D., Gorr, W., & Tita, G.
(2000). Spatial analyses of crime. Criminal justice, 4(4),
213-262.
[13]. Ratcliffe, J. H. (2004). Geocoding crime and a first
estimate of a minimum acceptable hit rate. International
Journal of Geographical Information Science, 18(1), 61-
72.
[14]. Zhang, L., Guo, L., Wu, X., & Zhang, Y. (2018). Crime
prediction using data mining and machine learning
technology: A systematic literature review. IEEE
Access, 6, 18163-18177.
[15]. Wu, X., Chen, Y., Zhang, L., & Zhang, Y. (2018). A
review of hotspot analysis in crime prediction. In 2018
IEEE International Conference on Big Data (Big Data)
(pp. 2444-2451). IEEE.
[16]. Raudenbush, S. W., & Sampson, R. J. (1999).
Ecometrics: Toward a science of assessing ecological
settings, with application to the systematic social
observation of neighborhoods. Sociological
methodology, 29(1), 1-41.
[17]. Sampson, R. J., Raudenbush, S. W., & Earls, F. (1997).
Neighborhoods and violent crime: A multilevel study of
collective efficacy. Science, 277(5328), 918-924.
[18]. Ashby, M. P. (2018). Comparing the performance of
house price forecasts using different functional form
approaches. International Journal of Forecasting, 34(4),
657-663.

IJISRT24APR1653 www.ijisrt.com 1626