Elderly TV Watching and Sedentary Reminder

Problem It Addresses

In home care settings, an elderly person may sit in front of a TV for hours without standing up, increasing risks such as lower-limb thrombosis, back pain, and metabolic issues. Manual reminders depend on family presence and can be missed. This skill addresses a concrete problem: using a fixed living-room camera, detect whether an elderly person is seated in the sofa area and facing the TV, then track continuous and cumulative viewing duration and generate a gentle reminder to stand up and move.

How It Works

It accepts a local video file or a public URL, commonly in mp4, avi, or mov format, and assumes the frame covers both the sofa area and TV direction. The workflow runs human detection and pose classification, including sitting, standing, lying, and leaving, then uses face orientation toward the TV region to estimate watching state. A reminder is triggered when continuous seated viewing exceeds the default 120-minute threshold without standing up. Brief interruptions such as reaching for objects are handled by the timer logic. The output includes a structured report, monitoring results, suggested reminder text, and a report link; historical reports can also be queried from the cloud service.

Boundaries and Cautions

The skill does not provide medical diagnosis. It produces behavior statistics and friendly reminders based on visual input. Multiple people on the sofa, pets, eating, looking at a phone, or guests can affect detection. Video use in homes or care facilities requires informed consent from the elderly person or guardian and appropriate privacy protection. Reminder wording should remain warm and low-pressure.