S.P.I.N.E.
CompletedSmart Posture Imaging & Notification Engine
FIU Senior Design Project · Team 8 · Spring 2026 · Completed
Completed camera-based posture monitoring system with real-time haptic feedback via a custom BLE wristband, backed by a containerized cloud service for session and posture data.
Last Updated: April 2026
Development Progress
Problem Statement
Poor posture during extended computer use leads to chronic pain, reduced productivity, and long-term health issues. While many people are aware of the importance of good posture, they often forget to self-correct during focused work.
Existing solutions rely on intrusive wearables or lack real-time feedback. S.P.I.N.E. provides a non-invasive, privacy-focused solution using computer vision and subtle haptic feedback to promote healthier sitting habits.
Solution Overview
Camera Detection
MediaPipe-based computer vision analyzes posture in real-time using existing webcams or dedicated cameras.
Wearable Feedback
Custom wristband with vibration motors and LED indicators provides gentle, non-intrusive posture correction reminders.
Desktop App & Cloud Backend
Desktop application for posture monitoring with a cloud backend for user management, session tracking, and posture data storage.
Technical Architecture
Computer Vision Module
Python-based application using MediaPipe Pose for real-time skeletal tracking. Analyzes neck angle, shoulder alignment, and spine curvature to detect slouching and forward head posture.
Wearable Device
ESP32-based wristband with Bluetooth connectivity. Receives posture alerts and triggers vibration patterns or LED sequences based on severity and user preferences.
Cloud Backend
A containerized backend service deployed to a personal VPS, handling user accounts, organization licensing, posture sessions, posture events, and aggregated stats with token-based access control.
Privacy Design
All posture analysis and processing happens locally on the user's device. No video or skeletal data leaves the device. Only posture status and user information are stored in the cloud for session tracking.
Technology Stack
Computer Vision
Hardware
Desktop Application
Backend & Cloud
Key Features
Real-time Posture Analysis
Continuous monitoring with sub-second latency
Personalized Profiles
ML-based calibration to individual body types
Multi-modal Feedback
Vibration patterns, LED alerts, and app notifications
Privacy-First Design
All processing local; only posture status stored in cloud for tracking
Goal Tracking
Historical data and progress visualization
Customizable Sensitivity
Adjust thresholds for different work scenarios
Target Users
Office Workers
Desk professionals spending 6+ hours daily at computers
Corporate Wellness
Companies seeking to reduce ergonomic injury claims
Team & Roles
Faculty Mentor: Dr. Constantinos Zekios, Florida International University
Collaborative team effort with cross-functional responsibilities in hardware design, software development, machine learning, and system integration.