Denoising and completion filters for human motion software: A survey with code
Enrico Martini, Andrea Calanca, Nicola Bombieri · Computer Science Review · 2025
Software platforms for human motion analysis are increasingly utilized across various fields, from Healthcare to Industry 5.0. However, the inherent inaccuracies of these platforms often lead to noisy observations of human poses or periods of missing information. As a result, data filtering for denoising or completion is a fundamental step before data analysis. Over the years, different techniques have been proposed, from general-purpose solutions based on low-pass filters to more advanced and embedded approaches based on state observers rather than deep learning. This survey presents the current state-of-the-art filtering solutions for denoising and completing data generated by software platforms for human motion analysis. It focuses on 3D positional data extrapolated through marker-based or marker-less motion capture systems. The survey proposes a concise taxonomy based on filter technology and application assumptions. For each class, it summarizes the basic concepts and reports application feedback collected from the literature. The survey also includes implementation codes or links to the authors’ original codes, enabling readers to quickly reproduce all the algorithms in different experimental settings ( https://github.com/PARCO-LAB/mocap-refinement ).