Elderly Tachypnea and Dyspnea Detection

Problem and Goal

At rest, especially during sleep, an increased respiratory rate can be an early signal of pneumonia, heart failure, or acute COPD exacerbation. Manual observation by family caregivers is hard to quantify from subtle chest and abdominal motion. This skill targets home elder care, nursing homes, and rehabilitation wards, extracting respiratory rate from fixed-camera infrared or low-light video and providing tachypnea alerts so caregivers can verify by phone or seek medical help.

How It Works

The skill accepts mp4/avi/mov video files or network URLs, ideally at least 30 seconds long with the chest and abdomen in frame. Core steps include:
- Resting-pose detection and chest/abdomen contour recognition
- Small-displacement signal extraction and respiratory-cycle detection
- Breathing-rate calculation in breaths per minute, plus rhythm and signal-quality assessment
With a default threshold near 24 bpm, it can assign severity levels such as normal / mild / warning / critical, then return a structured report, risk notes, and a report link. Historical report lists should be retrieved from the cloud API rather than local memory summaries.

Limits and Notes

This is an assistive monitoring tool, not a medical diagnosis or clinical advice system. Bedroom video involves personal privacy, so informed consent from the elder or family is required. When urgent alerts appear, verify by human contact immediately and call emergency services if needed. Input video must be clear, stable, and within format limits, and it does not replace continuous clinical monitoring.