Scholarlink

<p>Ergonomics is the scientific study of designing workplaces, systems, and tasks to fit the capabilities and limitations of the human body. It aims to optimize human well-being and overall performance by minimizing physical strain, improving posture, and reducing the risk of work-related musculoskeletal disorders. Ergonomics studies how individuals interact with their work environments, especially regarding posture and comfort. This research aims to observe and analyze various sitting postures commonly adopted by individuals working from home, particularly those aged between 20 and 35. The goal is to develop an automatic posture recognition system using transfer learning techniques based on Convolutional Neural Networks (CNNs). A real-time dataset of five common sitting postures—Slouching, Leaning Forward, Leaning Backward, Upright, and Crossed Legs—are collected and categorized into three groups. The system will identify key points in human posture images, classify the type of posture, and provide valuable inferences. By comparing the performance of four different transfer learning models, the study expects to determine the most accurate one for posture classification, with preliminary results suggesting that the Xception model offers the best performance. The outcome of this research contributes to improving ergonomic awareness and helps to reduce musculoskeletal issues among remote workers by providing personalized fitness recommendations.</p>

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

A diagnostic model for evaluating the posture recognition using multivariate Gaussian deep CNN for determining musculoskeletal disorders

  • 期刊论文

  • 发表日期：

    2025年9月5日

查看全文
  • A Sheik Abdullah,
  • R Parkavi,
  • S Geetha

摘要

Ergonomics is the scientific study of designing workplaces, systems, and tasks to fit the capabilities and limitations of the human body. It aims to optimize human well-being and overall performance by minimizing physical strain, improving posture, and reducing the risk of work-related musculoskeletal disorders. Ergonomics studies how individuals interact with their work environments, especially regarding posture and comfort. This research aims to observe and analyze various sitting postures commonly adopted by individuals working from home, particularly those aged between 20 and 35. The goal is to develop an automatic posture recognition system using transfer learning techniques based on Convolutional Neural Networks (CNNs). A real-time dataset of five common sitting postures—Slouching, Leaning Forward, Leaning Backward, Upright, and Crossed Legs—are collected and categorized into three groups. The system will identify key points in human posture images, classify the type of posture, and provide valuable inferences. By comparing the performance of four different transfer learning models, the study expects to determine the most accurate one for posture classification, with preliminary results suggesting that the Xception model offers the best performance. The outcome of this research contributes to improving ergonomic awareness and helps to reduce musculoskeletal issues among remote workers by providing personalized fitness recommendations.