Skip to content

Home

Jump to bottom
Ehsanul-karim edited this page Dec 20, 2024 · 4 revisions

Welcome to the CIS wiki!

Experience the transformative power of our Criminal Identification System, a revolutionary web-based platform that modernizes law enforcement with advanced AI capabilities, intuitive design, and unparalleled efficiency.

🌟 Team Members
Ehsanul Karim Ankon Chowdhury
Roll: 1907039 Roll: 1907048
Github Profile: Ehsan Github Profile: Ankon

Supervisor:

Dr. Sk. Md. Masudul Ahsan

Professor, Dept. of Computer Science and Engineering, KUET

🌟 Project Overview

Why CIS?

The Criminal Identification System (CIS) is a groundbreaking solution addressing the inefficiencies of traditional law enforcement processes. From the tedious sketching of suspects to the manual filing of charge sheets, CIS leverages cutting-edge technology to revolutionize the workflow, offering faster suspect identification, streamlined filing, and a robust database system.

Objectives:

  • Facial Recognition: Implement advanced Convolutional Neural Networks (CNNs) for accurate suspect identification.
  • Web Application: Provide a user-friendly platform for filing charge sheets and complaints online.
  • Data Storage: Maintain a secure and efficient database for FIRs, complaints, and suspect details.
  • Law Enforcement Empowerment: Equip law enforcement with tools for advanced search, record updates, and efficient case management.

Key Features:

  • Facial Feature Extraction: Powered by ResNet50, achieving up to 88% accuracy under favorable conditions. Capable to extract features like Hair Color, Hair Length, Hair Type, Skin Tone, Age Group, Gender from an image.
  • Web-Based Filing: Easy online submission of charge sheets and FIRs.
  • Advanced Search and Notification: Streamlined access to case records and real-time updates.
  • Role-Based Access: Tailored functionalities for Citizens, Police Officers, and Super Admins.
  • Scalable Architecture: Designed for adaptability with future technological advancements.

🌟 Features Showcase

Citizen Role:

  • Register and Login: Secure authentication.
  • File Complaints: Include suspect details and case descriptions.
  • Receive Notifications: Real-time updates on complaints.

Police Role:

  • Case Management: View and update FIRs.
  • Criminal Records: Add or modify suspect profiles.

Super Admin Role:

  • Advanced Management: Assign officers to specific jurisdictions, manage FAQs, and oversee the system.

🌟 System Design

Methodology:

  • Frontend: Built using Bootstrap, SCSS, and custom JavaScript for a modern, dynamic user experience.
  • Backend: Developed with Django to ensure smooth functionality and secure data handling.
  • Database: Utilized SQLite for lightweight and structured data storage.
  • AI Integration: Pre-trained ResNet50 for feature extraction, with custom layers for enhanced classification.

Tools and Libraries:

  • Frameworks: Django, TensorFlow, Keras
  • Libraries: OpenCV, Matplotlib
  • Development Tools: Jupyter Notebook, VS Code

🌟 Results and Implementation

Web Application:

  • User-Friendly Interface: Designed with seamless navigation.
  • Role-Based Dashboards: Tailored pages for citizens, police, and super admins.

AI Model:

  • ResNet50 Architecture: Integrated residual learning for accurate image classification.
  • Dataset: Over 17,000 high-quality images with diverse features such as gender, age group, skin tone, and more.

🌟 Socio-Economic and Ethical Impact

  • Community Engagement: Empowers citizens to participate actively in reporting crimes.
  • Enhanced Accessibility: Reduces barriers to justice by digitalizing complaint submission.
  • Data Security: Ensures privacy and confidentiality of sensitive information.
  • Scalability: Future-proof design, adaptable to courtroom integrations and broader applications.

🌟 Conclusion and Future Work

Achievements:

The CIS successfully integrates advanced AI models with a robust web platform, streamlining law enforcement processes and enhancing operational efficiency.

Future Enhancements:

  • Payment Gateway: For processing fines or legal fees.
  • Courtroom Integration: Extending functionalities to cover legal proceedings.
  • Dataset Expansion: Larger datasets for improved model accuracy.
  • Multi-Platform Availability: Mobile app integration for broader accessibility.

🌟 Pages

🌟 References

[1] Alarifi, J. S., Goyal, M., Davison, A. K., Dancey, D., Khan, R., & Yap, M. H. (2017). Facial Skin Classification Using Convolutional Neural Networks (pp. 479–485). https://doi.org/10.1007/978-3-319-59876-5_53

[2] LIEW, S. S., KHALIL-HANI, M., AHMAD RADZI, S., & BAKHTERI, R. (2016). Gender classification: a convolutional neural network approach. TURKISH JOURNAL OF ELECTRICAL ENGINEERING & COMPUTER SCIENCES, 24, 1248–1264. https://doi.org/10.3906/elk-1311-58

[3] Muhammad, U. R., Svanera, M., Leonardi, R., & Benini, S. (2018). Hair detection, segmentation, and hairstyle classification in the wild. Image and Vision Computing, 71, 25–37. https://doi.org/10.1016/j.imavis.2018.02.001

[4] Mustapha, M. F., Mohamad, N. M., Osman, G., & Ab Hamid, S. H. (2021). Age Group Classification using Convolutional Neural Network (CNN). Journal of Physics:ConferenceSeries,2084(1),012028.https://doi.org/10.1088/1742-6596/2084/1/012028

[5] Deep Learning with Python Book by Francois Chollet

[6] Django documentation | Django documentation | Django (djangoproject.com)

[7] [ResNet and ResNetV2 (keras.io)] (https://keras.io/api/applications/resnet/#resnet50-function)

[8] AI-Generated Faces: Diverse & Customizable | Generated.Photos

🌟 Acknowledgement

We sincerely thank to Dr. Sk. Md. Masudul Ahsan Sir for his guidance, inspiration, support, and encouragement in completing the Criminal Identification System project successfully

#Author's Ehsanul karim & Ankon Chowdhury

Clone this wiki locally