Scholarlink

In 2017, The World Health Organization (WHO) initiated a global campaign to improve rehabilitation services by 2030, focusing on the approximately 1 billion people who do not have access to proper care. In rehabilitation therapy, pose detection appears to be a promising tool, allowing for exact tracking and analysis of patients’ motions. By incorporating pose detection into therapy sessions, physicians can design personalized and interactive rehabilitation programs that are suited to each patient’s needs. This system provides real-time feedback and monitoring, allowing patients to do therapeutic activities with increased accuracy and efficacy. Machine Learning (ML) also plays an important role in healthcare by allowing for autonomous learning from previous data, which is required for predictive modelling and optimizing rehabilitation procedures. This work compares 16 different classical machine learning models in which for classifying diabetic frozen shoulders which are Flexion, Abduction and External rotation where Random Forest Classifier was the highest model where it achieved 92% indicating the efficacy of the proposed approach. Future study will involve merging segmentation approaches and multi-modal data to improve model robustness and accessibility and enhancing rehabilitation outcomes and patient quality of life worldwide.

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

Comparative Study: Using Machine Learning Models Based Rehabilitation Therapy to Classify Diabetic Frozen Shoulder Exercises

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Ziad A. Ahmed,
  • Mohamed Sherif,
  • Maha Abdelmohsen,
  • Omnia Saeed Mahmoud Ahmed,
  • Asmaa M. Al-Emrany

摘要

In 2017, The World Health Organization (WHO) initiated a global campaign to improve rehabilitation services by 2030, focusing on the approximately 1 billion people who do not have access to proper care. In rehabilitation therapy, pose detection appears to be a promising tool, allowing for exact tracking and analysis of patients’ motions. By incorporating pose detection into therapy sessions, physicians can design personalized and interactive rehabilitation programs that are suited to each patient’s needs. This system provides real-time feedback and monitoring, allowing patients to do therapeutic activities with increased accuracy and efficacy. Machine Learning (ML) also plays an important role in healthcare by allowing for autonomous learning from previous data, which is required for predictive modelling and optimizing rehabilitation procedures. This work compares 16 different classical machine learning models in which for classifying diabetic frozen shoulders which are Flexion, Abduction and External rotation where Random Forest Classifier was the highest model where it achieved 92% indicating the efficacy of the proposed approach. Future study will involve merging segmentation approaches and multi-modal data to improve model robustness and accessibility and enhancing rehabilitation outcomes and patient quality of life worldwide.