APPLIED_RESEARCH_PAPER

Multimodal Posture Monitoring

Computer Vision • IoT Sensor Fusion • Healthcare

Springer_Link

Abstract

Sedentary lifestyles are a silent health crisis. This research proposes a multimodal system combining ultrasonic sensors and computer vision to monitor posture and eye-screen distance in real-time, preventing musculoskeletal disorders.

PUBLICATION: Smart Systems: Innovations in Computing (Springer)

Sensor Fusion Pipeline

DATA INGESTION: MQTT / OPENCV

1. Visual Stream (OpenCV)

Detects facial landmarks to estimate head tilt and eye-to-screen distance vectors. Optimized for low-compute environments (Raspberry Pi 4).

2. IoT Sensor Array

Ultrasonic distance sensors verify visual data. Fusion logic (Kalman Filter) reconciles discrepancies between visual estimates and physical sensor readings.

Performance_Data

94.2%

Accuracy

<200ms

Inference Latency

Real-time

Alert Generation

"Bridging the gap between hardware sensors and computer vision to create ambient health monitoring systems."