Scholarlink

<p>In recent years, the study of human gait, a fundamental daily behavior, known as gait analysis, has gained significant attention. While gait recognition has been somewhat overshadowed by pattern recognition, it offers distinct advantages like non-intrusiveness and independence from subject cooperation. This study explores the integration of FAST and Harris Corner detection methods with classifiers such as MLP, Random Forest, and extreme gradient boosting using the CASIA-A dataset. The primary goal was to enhance accuracy through this combined approach. Results indicate that the extreme gradient boosting classifier achieved the highest accuracy at 92.77%, surpassing MLP and Random Forest. Key performance metrics evaluated included recognition accuracy, area under the curve, and root mean squared error, demonstrating the effectiveness of combining corner detection algorithms with advanced classifiers in enhancing gait recognition systems.</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

CASIA-A Gait: Human Identification Through Gait Analysis With FAST and Harris Corner Detector Algorithms

  • 期刊论文

  • 发表日期：

    2025年3月28日

查看全文
  • Junainah Abd Hamid,
  • K. N. Raja Praveen,
  • Amandeep Kaur,
  • Mandeep Kaur Chohan,
  • Shivakrishna Dasi,
  • P. P. N. G. Phani Kumar,
  • Devendra Singh,
  • Ahmed Alkhayyat

摘要

In recent years, the study of human gait, a fundamental daily behavior, known as gait analysis, has gained significant attention. While gait recognition has been somewhat overshadowed by pattern recognition, it offers distinct advantages like non-intrusiveness and independence from subject cooperation. This study explores the integration of FAST and Harris Corner detection methods with classifiers such as MLP, Random Forest, and extreme gradient boosting using the CASIA-A dataset. The primary goal was to enhance accuracy through this combined approach. Results indicate that the extreme gradient boosting classifier achieved the highest accuracy at 92.77%, surpassing MLP and Random Forest. Key performance metrics evaluated included recognition accuracy, area under the curve, and root mean squared error, demonstrating the effectiveness of combining corner detection algorithms with advanced classifiers in enhancing gait recognition systems.