Scholarlink

This study explores the utilization of marker less object tracking combined with machine learning (ML) techniques to analyze drawing patterns in Parkinson's disease (PD) patients. Motor symptoms, such as tremors and bradykinesia, are prominent features of PD and often help in its identification. Diagnosing PD accurately is crucial yet complex, with error rates in clinical practice. Diagnostic tools, including genetic testing and imaging, help but still have limitations. Drawing tasks, particularly using the Archimedes spiral, have emerged as potential diagnostic tools. Recent advancements in ML, particularly in computer vision techniques like DeepLabCut (DLC), offer new avenues for PD diagnosis. By tracking object positions and analyzing drawing characteristics, ML algorithms can potentially identify patterns indicative of PD. This study aims to leverage DLC and ML algorithms to analyze drawing data from PD patients and healthy individuals. Preliminary findings suggest promising results in classifying PD patients based on drawing patterns. The implications of this research extend to clinical practice, highlighting the potential of ML and computer vision tools as complementary diagnostics for PD.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Patterns in Drawings of Parkinson’s Disease Patients Versus Healthy People Utilizing Markerless Object Tracking with Machine Learning

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • José Renato Munari Nardo,
  • Caio Tonus Ribeiro,
  • Camille Marques Alves,
  • Daniel Hilário Silva,
  • Eduardo Moura Neto,
  • Luanne Cardoso Mendes,
  • Adriano Alves Pereira,
  • Adriano de Oliveira Andrade

摘要

This study explores the utilization of marker less object tracking combined with machine learning (ML) techniques to analyze drawing patterns in Parkinson's disease (PD) patients. Motor symptoms, such as tremors and bradykinesia, are prominent features of PD and often help in its identification. Diagnosing PD accurately is crucial yet complex, with error rates in clinical practice. Diagnostic tools, including genetic testing and imaging, help but still have limitations. Drawing tasks, particularly using the Archimedes spiral, have emerged as potential diagnostic tools. Recent advancements in ML, particularly in computer vision techniques like DeepLabCut (DLC), offer new avenues for PD diagnosis. By tracking object positions and analyzing drawing characteristics, ML algorithms can potentially identify patterns indicative of PD. This study aims to leverage DLC and ML algorithms to analyze drawing data from PD patients and healthy individuals. Preliminary findings suggest promising results in classifying PD patients based on drawing patterns. The implications of this research extend to clinical practice, highlighting the potential of ML and computer vision tools as complementary diagnostics for PD.