Automated gap-filling for marker-based biomechanical motion capture data

A Fully-Automatic Gap Filling Approach for Motion Capture Trajectories

“…These methods are grounded on the idea that some markers' trajectories may be highly related due to bone length and motion range constraints. The pattern fill method uses a trajectory without gaps (the donor trajectory) to fill the selected gap, whereas the rigid body fill uses at least three donor trajectories that are assumed to be all part of the same rigid body [5,16]. A full description of these methods can be found in [5] and [17].…”

“…The pattern fill method uses a trajectory without gaps (the donor trajectory) to fill the selected gap, whereas the rigid body fill uses at least three donor trajectories that are assumed to be all part of the same rigid body [5,16]. A full description of these methods can be found in [5] and [17]. Skeleton-based methods require prior knowledge about skeleton constraints and are highly dependent on how accurately these constraints can be represented.…”

“…Existing methods cannot satisfy all requirements for a generic and universal gap filling approach. To overcome the need of manually selecting the most appropriate method, Camargo et al [5] proposes an automatic pipeline, where spline interpolation, pattern fill and rigid body fill are iteratively selected. Despite replacing manual gap filling, the method may still fail when rigid body assumptions cannot be considered.…”

“…This phenomenon is quite frequent and translates into incomplete marker trajectories that can occur due to a marker falling off or due to temporary marker occlusion caused by other body segments, interaction with physical objects and overall non-optimal experimental setups. The resultant gaps impair data quality and compromise the accuracy of the analysis [5][6][7].…”

“…Some of them are available in commercial MoCap post-processing software such as those provided by leading MoCap companies, such as Qualysis (Qualysis AB, Gothernburg, Sweden) or Vicon (Oxford Metric, Limited, Oxford, England). Even though they provide satisfactory results in many situations, they still rely on manual intervention where the user needs to visually inspect each gap and decide how it should be filled [5]. This dependency on the user results in a time-consuming and cumbersome process.…”