Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Data Acquisition Through Participatory Design for Automated Rehabilitation Assessment

Tamim Ahmed, Zhaoyi Guo, Mohammod Shaikh Sadid Khan, Thanassis Rikakis, Aisling Kelliher

TL;DR

The paper addresses the challenge of automated ARAT assessment by employing a participatory design approach to create an end-to-end data collection and labeling workflow. It introduces a low-cost, four-camera capture system, a standardized ARAT segmentation vocabulary with a React-based tool, and an integrated rating interface to generate clinician-compatible annotations. The study reports data from 106 ARAT sessions (1800 task videos) with 760 videos segmented into ~3000 segments and rated by clinicians, demonstrating high usability and alignment with clinical practice (roughly 90% of segmented videos rated, segmentation errors <4%). The resulting data pipeline and interfaces enable training AI tools for automated segmentation and rating, with potential generalization to other domains requiring expert movement assessment and augmented intelligence in rehabilitation.

Abstract

Through participatory design, we are developing a computational system for the semi-automated assessment of the Action Research Arm Test (ARAT) for stroke rehabilitation. During rehabilitation assessment, clinicians rate movement segments and components in the context of overall task performance. Clinicians change viewing angles to assess particular components. Through studies with clinicians, we develop a system that includes: a) unobtrusive multi-camera capture, b) a segmentation interface for non-expert segmentors, and c) a rating interface for expert clinicians. Five clinicians independently captured 1800 stroke survivor videos with <5$\%$ errors. Three segmentors have segmented 760 of these videos, averaging 20 seconds per segment. They favor the recommended camera view $>$ 90\%. Multiple clinicians have rated the segmented videos while reporting minimal problems. The complete data will be used for training an automated segmentation and rating system that empowers the clinicians as the ratings will be compatible with clinical practice and intuition.

Data Acquisition Through Participatory Design for Automated Rehabilitation Assessment

TL;DR

The paper addresses the challenge of automated ARAT assessment by employing a participatory design approach to create an end-to-end data collection and labeling workflow. It introduces a low-cost, four-camera capture system, a standardized ARAT segmentation vocabulary with a React-based tool, and an integrated rating interface to generate clinician-compatible annotations. The study reports data from 106 ARAT sessions (1800 task videos) with 760 videos segmented into ~3000 segments and rated by clinicians, demonstrating high usability and alignment with clinical practice (roughly 90% of segmented videos rated, segmentation errors <4%). The resulting data pipeline and interfaces enable training AI tools for automated segmentation and rating, with potential generalization to other domains requiring expert movement assessment and augmented intelligence in rehabilitation.

Abstract

Through participatory design, we are developing a computational system for the semi-automated assessment of the Action Research Arm Test (ARAT) for stroke rehabilitation. During rehabilitation assessment, clinicians rate movement segments and components in the context of overall task performance. Clinicians change viewing angles to assess particular components. Through studies with clinicians, we develop a system that includes: a) unobtrusive multi-camera capture, b) a segmentation interface for non-expert segmentors, and c) a rating interface for expert clinicians. Five clinicians independently captured 1800 stroke survivor videos with <5 errors. Three segmentors have segmented 760 of these videos, averaging 20 seconds per segment. They favor the recommended camera view 90\%. Multiple clinicians have rated the segmented videos while reporting minimal problems. The complete data will be used for training an automated segmentation and rating system that empowers the clinicians as the ratings will be compatible with clinical practice and intuition.
Paper Structure (10 sections, 6 figures, 3 tables)

This paper contains 10 sections, 6 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Methods
  3. Standardized Capturing Setup for ARAT Exercises
  4. ARAT Segmentation Vocabulary and Standardized Segmentation Tool
  5. Rating Interface
  6. Results
  7. Data Acquisition Summary
  8. Segmentation Interface Usability
  9. Rating summary
  10. Discussion and Conclusion

Figures (6)

  • Figure 1: (A) The multi-camera ARAT exercise capture setup with the (B) ipsilateral, (C) transverse, and (D) contralateral view.
  • Figure 2: Calibration screen, camera check screen, and ARAT administration screen of our custom-developed capture interface
  • Figure 3: Segmentation interface and its key components
  • Figure 4: Rating interface and its key components
  • Figure 5: This plot demonstrates the decrease in segmentation time with the number of segmented videos for 3 segmentors as their experience with the interface increases. On the y-axis, we have the average time taken to complete 10 segments. Colored lines indicate different types of movement segments.
  • ...and 1 more figures