AI Physical Function Assessment System Using Convolutional Neural Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional physical function assessments require trained clinicians, leading to inter-rater variability, increased healthcare costs, and delayed assessments in resource-constrained settings, and are not sustainable for the growing aging population.
Innovation Solution
An AI-based physical assessment system using a mobile device with a camera for capturing video frames, employing convolutional neural networks to track Person of Interest and Object of Interest, extract features, and recognize postures and gait, enabling automated and standardized physical function assessments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional observation-based physical function assessments are conducted by trained clinicians, then measurement precision can be maintained through expert judgment, but device complexity and healthcare costs increase due to requiring trained personnel
Solution Approach 1:
The patent replaces the mechanical/observational assessment system performed by clinicians with an automated AI-based computer vision system. The system uses cameras to capture video frames and employs convolutional neural networks to automatically detect body key points, recognize postures, and assess physical function, eliminating the need for trained clinicians to perform observations while maintaining or improving measurement precision through objective algorithmic analysis
Solution Approach 2:
The assessment system performs self-assessment by automatically capturing video data, processing it through AI algorithms, and generating physical function assessments without requiring external clinician intervention. The system serves itself by integrating all assessment functions into an automated workflow that operates independently of trained personnel
2Adaptability or versatility
If traditional physical function assessments are conducted manually by clinicians, then adaptability to individual patient needs can be achieved through professional judgment, but productivity decreases due to time-consuming observation and interpretation processes
Solution Approach 1:
The system enables continuous assessment by automatically processing video frames in real-time or near-real-time. The AI algorithms continuously analyze body key points, postures, and movements without interruption, eliminating the gaps and delays inherent in manual clinician observations. This continuous automated processing significantly increases productivity while maintaining adaptability through algorithmic adjustment to different assessment protocols and patient characteristics
3Reliability
If observation-based physical function assessments are performed by multiple clinicians, then comprehensive evaluation can be achieved through multiple perspectives, but inter-rater variability increases leading to inconsistent results
Solution Approach 1:
The patent applies homogeneity by using a standardized AI-based assessment algorithm that treats all patients uniformly. The same convolutional neural network model and body key point detection methodology are applied consistently across all assessments, eliminating the variability introduced by different clinicians' observation techniques, interpretations, and levels of training. This homogeneous approach ensures reliable and consistent results across diverse patient populations and assessment settings
Data Source
AI summary
Various embodiments of devices, systems, and methods for providing AI-based physical function assessment recordings and assessment performance analytics for a subject are described. A series of video frames are obtained that include the subject. Computer vision techniques that use artificial neural networks may be applied to the video frames to: detect a Person of Interest (POI) and an Object of Interest (OOI) in the video frames; track movement of the POI and the location of the OOI in subsequent video frames; detect body key points; and detect postures and posture transitions of the POI. Physical function indicators may be calculated and function analytics provided based on the assessment.