MISO: Monitoring Inactivity of Single Older Adults at Home using RGB-D Technology

TL;DR

This work tackles the problem of monitoring inactivity in a single older adult’s home without compromising privacy. It introduces MISO, an on-device RGB-D pipeline that combines depth-based foreground extraction, color-based motion detection, and pet-motion suppression to detect inactivity events defined as movement-free intervals longer than $1$ second, logging their times and durations. Activity patterns are modeled with a Poisson process for movements and an exponential distribution for inter-movement intervals, with a maximum-likelihood rate $lambda_{MLE} = 1 / mean(x)$. Across lab and field tests, the system shows high accuracy, robustness to dim lighting and TV flicker, and outperforms state-of-the-art methods for small motions, all while preserving privacy by avoiding image storage and internet connectivity. The approach supports long-term, non-intrusive home monitoring and offers a foundation for personalized health insights and timely interventions, with future work addressing very dark environments and internet-enabled emergency capabilities.

Abstract

A new application for real-time monitoring of the lack of movement in older adults' own homes is proposed, aiming to support people's lives and independence in their later years. A lightweight camera monitoring system, based on an RGB-D camera and a compact computer processor, was developed and piloted in community homes to observe the daily behavior of older adults. Instances of body inactivity were detected in everyday scenarios anonymously and unobtrusively. These events can be explained at a higher level, such as a loss of consciousness or physiological deterioration. The accuracy of the inactivity monitoring system is assessed, and statistics of inactivity events related to the daily behavior of older adults are provided. The results demonstrate that our method achieves high accuracy in inactivity detection across various environments and camera views. It outperforms existing state-of-the-art vision-based models in challenging conditions like dim room lighting and TV flickering. However, the proposed method does require some ambient light to function effectively.

Figures (11)

• Figure 1: Real-time system for monitoring older adults in common home scenarios (inactivity monitoring of routine seating areas). Anonymous monitoring uses a compact system consisting of an RGB-D camera and a small computer processor. Anonymity is maintained by discarding captured images after processing.
• Figure 2: RGB value difference of one color pixel (selected on the human chest) in consecutive frames under the following conditions: (a) no human motion under constant low light; (b) same environment, without human motion but with TV light changes; (c) human movement only. (d) The threshold for distinguishing noise (a) from true body motion (c), see main text for details. (Scene Luma Rec. 601 rec601$Y'$ is 20.5).
• Figure 3: A person sitting on a couch. The grey area shows the foreground detected using the depth map (a) before the region growing and (b) after the region growing. Real human motion (red) is not detected in the incomplete foreground before the region growing.
• Figure 4: Object detector helps remove pet motion. (a) Pet movement occurs while humans are inactive. (b) The pet is successfully detected, and the relative region is removed from the foreground, allowing the human inactivity count to continue. (c) The pet is not detected by the object detector. (d) Pet motion is not removed, resulting in the cessation of the human inactivity count.
• Figure 5: Examples of inactivity detection evaluation in different scenarios (front view in office, lab, and homes). The top row shows people in motion, and the bottom row shows people remaining inactive. Detected humans are marked with blue boxes, and areas with detected motion are highlighted in red pixels.