Skip to content

HarshBang/CrimeIntel-Companion

Repository files navigation

logo

AI-Powered Crime Scene Assistant App

A mobile Android application designed to assist law enforcement officers directly at crime scenes, especially in regions lacking timely forensic support. The app aims to improve conviction rates in minor cases by enabling real-time evidence capture, secure storage, AI assistance, and SOP guidance.

Android Firebase AWS Amazon S3 PyTorch OpenAI Python

🎯 Project Goals

  • Support law enforcement officers during on-site crime scene documentation
  • Improve traceability, accuracy, and handling of crime scene evidence
  • Minimize delays caused by forensic expert shortages or lab backlogs

Key Features (So Far)

Feature Status
Authentication with Firebase ✅ Implemented
Geo-tagging with device GPS ✅ Implemented
Upload to AWS S3 (with IAM & Cognito) ✅ Implemented
Object Detection (Faster R-CNN model with 14 classes) 🔄 Model built, app integration pending
In-app Chatbot (OpenAI Generative API) ✅ Initial version done
SOPs PDF-based Assistance (via chatbot) 🔄 In progress
Offline-ready LLM + RAG Model for SOP Q&A 🔜 Planned

🧠 Evidence Detection Model (Coming Soon)

A custom Faster R-CNN model trained on 4,463 images across 14 evidence classes, such as:

  • Weapons
  • Blood & Fingerprints
  • Human hand & Body
  • Hair, Rope, Shoe-print, etc.

🔗 Model Repo – CIC-Evidence-Detection-RCNN-Prototype

📸 Demo & Previews

Interface Description
Main screen
Geotagging screen
Evidence Detection

🚧 Roadmap

  • Integrate object detection in-app (via lightweight or server-hosted model)
  • Implement SOP-based RAG Q&A using PDFs
  • Add offline LLM (Phi-2, Mistral, etc.) for edge environments
  • Create dashboards with role-based access
  • Case-wise evidence log with map and media view

👤 About Me

I am a passionate student at NMIMS School of Technology Management and Engineering, persuing Computer Science Engineering specializing in Data Science. CrimeIntel Companion reflects my drive to solve real-world societal challenges — particularly in the public safety and law enforcement space — through technology that can have real impact in the field.

This project combines my interests in AI, mobile app development, and cloud infrastructure to address critical gaps in crime scene documentation and evidence handling in underserved regions.

🤝 Acknowledgements

  • Dr. Naresh Vurukonda, Dr. Bhanushree Yalamanchili, Prof Vinayak Mukkawar – Project Mentors
  • Telangana Police SOP PDFs
  • Forensic Science & Criminal Psychology Podcast – Priyanshi Jain

📩 Contact Information

Feel free to connect for collaborations, feedback, or inquiries:

License: All rights reserved. The code is intended solely for academic purposes.

About

AI-powered Android app for crime scene image capture, geotagging, and forensic assistance.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages