ICU Patient Mobility Measurement Using RGB-D Sensors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for monitoring patient mobility in ICU settings are labor-intensive, prone to recall bias, and limited by the need for manual observation, which restricts data collection and accuracy, especially in complex environments with significant occlusions and pose variations.
Innovation Solution
A method using RGB-D sensors to analyze individual images, locate and identify patients, determine their pose, measure motion, and infer mobility levels by generating volumetric representations and heatmaps, classifying poses into discrete categories, and assigning numerical mobility values, employing a combination of machine learning and computer vision techniques.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual observation by nurses is used to monitor patient mobility, then mobility status can be recorded, but the process is highly impractical to scale and labor-intensive
Solution Approach 1:
The patent replaces manual mechanical observation by nurses with an automated computer vision system using RGB-D sensors and machine learning algorithms. The system automatically detects patient poses, tracks mobility events, and generates mobility scores without requiring human observers, thereby scaling productivity while maintaining measurement precision through continuous automated monitoring.
Solution Approach 2:
The system enables self-service by allowing the patient's mobility to be automatically monitored and recorded without requiring active participation from healthcare staff. The computer vision system independently performs detection, tracking, and analysis, making the measurement process autonomous and eliminating the need for manual intervention in data collection.
2Loss of time
If discrete subjective recordings of maximal mobility level are used, then mobility status is captured at specific time points, but the measurements are subject to recall bias and not truly representative of overall mobility level
Solution Approach 1:
The patent implements continuous monitoring through the computer vision system that operates throughout the observation period, capturing patient mobility at all times rather than at discrete intervals. This continuous action eliminates gaps in data collection, prevents recall bias by recording actual behavior in real-time, and provides a comprehensive representation of overall mobility levels through aggregated continuous data.
3Extent of automation
If non-invasive low-cost camera systems are used, then automation is achieved, but significant occlusions and pose variations limit detection accuracy
Solution Approach 1:
The patent transitions from traditional 2D camera systems to 3D RGB-D sensing that captures depth information, normal vectors, and spatial relationships. This dimensional enhancement allows the system to accurately detect patients even under occlusions by understanding three-dimensional body configurations and spatial context, thereby maintaining high detection accuracy while achieving full automation.
Solution Approach 2:
The system changes the parameter space by incorporating multiple features beyond simple 2D image data, including depth values, surface normals, and skeletal joint positions. These additional parameters provide robustness against occlusions and pose variations, enabling accurate patient detection and pose estimation in challenging environments while maintaining automation.
4Measurement precision
If direct observation by trained researchers is used, then comprehensive descriptive data sets are obtained with high accuracy, but the process is labor-intensive and limits the amount and duration of data collection
Solution Approach 1:
The patent replaces labor-intensive manual observation by trained researchers with an automated computer vision system that can continuously monitor patients without fatigue. This substitution enables collection of large volumes of mobility data over extended periods, maintaining high measurement accuracy through sophisticated algorithms while eliminating the productivity constraints imposed by human observers.
Data Source
AI summary
An embodiment in accordance with the present invention includes a technology to continuously measure patient mobility automatically, using sensors that capture color and depth images along with algorithms that process the data and analyze the activities of the patients and providers to assess the highest level of mobility of the patient. An algorithm according to the present invention employs the following five steps: 1) analyze individual images to locate the regions containing every person in the scene (Person Localization), 2) for each person region, assign an identity to distinguish ‘patient’ vs. ‘not patient’ (Patient Identification), 3) determine the pose of the patient, with the help of contextual information (Patient Pose Classification and Context Detection), 4) measure the degree of motion of the patient (Motion Analysis), and 5) infer the highest mobility level of the patient using the combination of pose and motion characteristics (Mobility Classification).