A review of 3D human pose estimation algorithms for markerless motion capture
This review helps researchers and practitioners in fields like robotics and health sciences make informed choices among proliferating methods, but it is incremental as it synthesizes existing work without introducing new algorithms.
The paper reviews leading 3D human pose estimation methods from the past five years, highlighting that modern techniques achieve an average error per joint of 20 mm, and proposes a taxonomy based on accuracy, speed, and robustness to classify methods and guide future research.
Human pose estimation is a very active research field, stimulated by its important applications in robotics, entertainment or health and sports sciences, among others. Advances in convolutional networks triggered noticeable improvements in 2D pose estimation, leading modern 3D markerless motion capture techniques to an average error per joint of 20 mm. However, with the proliferation of methods, it is becoming increasingly difficult to make an informed choice. Here, we review the leading human pose estimation methods of the past five years, focusing on metrics, benchmarks and method structures. We propose a taxonomy based on accuracy, speed and robustness that we use to classify de methods and derive directions for future research.