Computer vision in the field
I applied AI to helmet and seatbelt detection for UP Police, connecting model development with a concrete traffic-enforcement use case.
All workPROJECT NOTES / 2025
AI & SECURITYSurakshaNet.
Computer vision for a concrete enforcement problem.
SOFTWARE / SYSTEMS / CONNECTIONSSurakshaNetCONCEPTUAL PRODUCT MAP
The problem behind
the product.
I developed an AI-powered traffic enforcement system for UP Police, using computer vision to detect helmet and seatbelt violations in real time.
The challenge
Build a traffic-enforcement system that improves detection accuracy while making its infrastructure more economical to operate.
WHAT I WORKED ON
I used PyTorch and OpenCV for computer vision, with AWS as part of the project stack. My work focused on helmet and seatbelt detection and the practical tradeoff between accuracy and infrastructure cost.
Where the choices
matter.
Outcome scope
The quantified outcomes above are recorded in my résumé. The public repository demonstrates the detection implementation; it does not include the original comparison dataset, accuracy baseline, or infrastructure-cost breakdown.
The engineering stack
I brought PyTorch, OpenCV, and AWS together to connect applied machine learning with the infrastructure it needs to run.
What the evidence
supports.
40%
Résumé-reported accuracy gain
90%
Résumé-reported cost reduction
These outcomes are reported in my résumé. The public implementation does not include the comparison dataset or original cost breakdown.
Scope & limitations
Published as UP Police AI Traffic Surveillance, with helmet and seatbelt detection for Sitapur Police.
EXPLORE THE EVIDENCEView public code