Scholarlink

Pose estimation plays a pivotal role in robotics, augmented reality (AR), and human-computer interaction. This survey presents a comprehensive review of recent developments in 2D and 3D human and hand pose estimation techniques. We categorize the literature based on monocular, multi-view, depth-based, and mesh-based methods, highlighting the advantages and trade-offs of each. We also examine domain-specific applications in healthcare, telemedicine, and robotics, and identify major challenges such as occlusion, domain adaptation, and real-time processing. Finally, we discuss future directions, including self-supervised learning, sensor fusion, and explainability. This work aims to guide future research and inform the development of robust, generalizable pose estimation systems.

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 Comprehensive Survey of Computer Vision-Based Pose Estimation for Machine Learning and Deep Learning Approaches

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Mohamed Hadid,
  • Muhammad Abid

摘要

Pose estimation plays a pivotal role in robotics, augmented reality (AR), and human-computer interaction. This survey presents a comprehensive review of recent developments in 2D and 3D human and hand pose estimation techniques. We categorize the literature based on monocular, multi-view, depth-based, and mesh-based methods, highlighting the advantages and trade-offs of each. We also examine domain-specific applications in healthcare, telemedicine, and robotics, and identify major challenges such as occlusion, domain adaptation, and real-time processing. Finally, we discuss future directions, including self-supervised learning, sensor fusion, and explainability. This work aims to guide future research and inform the development of robust, generalizable pose estimation systems.