Work / SafeX

SafeX

AI computer-vision security system for face recognition, camera integration, attendance support and alert workflows.

SafeX is a real-time computer-vision security project focused on identity recognition, camera-based monitoring, attendance integration and educational use. The project explores data registration, model training, optimized recognition modes, alerting and practical deployment with company camera systems.

Links
GitHub

This project has stopped operating and is now preserved as an academic reference. It is no longer connected to any commercial activity.

Overview
SafeX was developed as an applied AI vision system that connects research, training data preparation, recognition algorithms and practical security/attendance workflows.
Problem
Small security deployments need identity recognition and attendance support without a heavy surveillance platform. The main challenge is keeping recognition accurate, fast and practical on normal camera infrastructure.
Role
Founder / Computer Vision Engineer
Approach
The system uses Python, Dlib face encodings, OpenCV processing, registration/training flows, optimized recognition modes, Firebase/Sheets integrations and RTSP/IP camera input.
Outcomes
Company
integrated with camera and attendance workflow
University
used for research and teaching
AI Vision
Dlib/OpenCV recognition pipeline

Project Highlights

Alert workflow

Security event notification and monitoring screen.

Registration flow

Face registration and training data management flow.

Main dashboard

Recognition dashboard for camera-based monitoring.

Video & Walkthrough

Timeline

2023
01. Discovery
2023
02. Architecture
2023
03. Prototype
2023
04. Validation
2023
05. Archive / Iterate

Behind The Project

Overview

SafeX was developed as an applied AI vision system that connects research, training data preparation, recognition algorithms and practical security/attendance workflows.

Problem

Small security deployments need identity recognition and attendance support without a heavy surveillance platform. The main challenge is keeping recognition accurate, fast and practical on normal camera infrastructure.

Approach

The system uses Python, Dlib face encodings, OpenCV processing, registration/training flows, optimized recognition modes, Firebase/Sheets integrations and RTSP/IP camera input.

Gallery

Interested in this project?

Let's build practical edge and AI systems for your workflow.

Contact
Share this project: