Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

AI WALKUP: A Computer-Vision Approach to Quantifying MDS-UPDRS in Parkinson's Disease

Xiang Xiang, Zihan Zhang, Jing Ma, Yao Deng

TL;DR

The paper addresses subjectivity and resource intensity in the MDS-UPDRS motor assessment for Parkinson's disease by proposing a smartphone-based computer-vision pipeline that derives six motor-item signals from monocular videos using MediaPipe pose and hand keypoints, followed by time-series feature extraction with tsfresh. The end-to-end framework runs on-device video capture and cloud-based pose analysis to provide objective movement descriptors and visualizations for clinicians. It demonstrates feasibility with data collected from healthy subjects and an Android app implementation, with plans to extend to PD patients and conduct clinical validation. The approach promises scalable, at-home motor assessment that can reduce clinic visits and clinician workload while enabling rapid, objective monitoring of disease progression.

Abstract

Parkinson's Disease (PD) is the second most common neurodegenerative disorder. The existing assessment method for PD is usually the Movement Disorder Society - Unified Parkinson's Disease Rating Scale (MDS-UPDRS) to assess the severity of various types of motor symptoms and disease progression. However, manual assessment suffers from high subjectivity, lack of consistency, and high cost and low efficiency of manual communication. We want to use a computer vision based solution to capture human pose images based on a camera, reconstruct and perform motion analysis using algorithms, and extract the features of the amount of motion through feature engineering. The proposed approach can be deployed on different smartphones, and the video recording and artificial intelligence analysis can be done quickly and easily through our APP.

AI WALKUP: A Computer-Vision Approach to Quantifying MDS-UPDRS in Parkinson's Disease

TL;DR

The paper addresses subjectivity and resource intensity in the MDS-UPDRS motor assessment for Parkinson's disease by proposing a smartphone-based computer-vision pipeline that derives six motor-item signals from monocular videos using MediaPipe pose and hand keypoints, followed by time-series feature extraction with tsfresh. The end-to-end framework runs on-device video capture and cloud-based pose analysis to provide objective movement descriptors and visualizations for clinicians. It demonstrates feasibility with data collected from healthy subjects and an Android app implementation, with plans to extend to PD patients and conduct clinical validation. The approach promises scalable, at-home motor assessment that can reduce clinic visits and clinician workload while enabling rapid, objective monitoring of disease progression.

Abstract

Parkinson's Disease (PD) is the second most common neurodegenerative disorder. The existing assessment method for PD is usually the Movement Disorder Society - Unified Parkinson's Disease Rating Scale (MDS-UPDRS) to assess the severity of various types of motor symptoms and disease progression. However, manual assessment suffers from high subjectivity, lack of consistency, and high cost and low efficiency of manual communication. We want to use a computer vision based solution to capture human pose images based on a camera, reconstruct and perform motion analysis using algorithms, and extract the features of the amount of motion through feature engineering. The proposed approach can be deployed on different smartphones, and the video recording and artificial intelligence analysis can be done quickly and easily through our APP.
Paper Structure (12 sections, 5 equations, 9 figures, 1 table)

This paper contains 12 sections, 5 equations, 9 figures, 1 table.

Table of Contents

  1. Introduction
  2. Related Work
  3. Methodology
  4. Data
  5. Body keypoints
  6. Hand keypoints
  7. Signals
  8. Feature extraction
  9. Experiment and Results
  10. Visualization of inputting signals
  11. The APP and demo results
  12. Conclusion

Figures (9)

  • Figure 1: The flowchart of the APP. The users record their specific action videos.
  • Figure 2: The pipeline of our method. Our method can be divided into user side and cloud side. The user can record the action video on APP and upload the videos. The cloud can analyze the videos and return the results.
  • Figure 3: The key points of the human body.
  • Figure 4: The key points of human hands.
  • Figure 5: Visualization of action two video.
  • ...and 4 more figures