Database generation for markerless tracking based on Deep Learning networks

Nicolau, Ana  Virginia Ruescas; Bernabé, Eduardo Parrilla; Ripoll, E. Medina; Jaén, D. Garrido; Mut, S. Alemany

doi:10.1016/j.gaitpost.2020.08.050

Cited by 2 publications

(3 citation statements)

References 1 publication

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A portion of the dataset “Human tracking dataset of 3D anatomical landmarks and pose keypoints” [ 39 ] has been used. Section 2.1.1 briefly introduces the workflow followed to obtain the whole dataset.…”

Section: Methodsmentioning

confidence: 99%

“…Further details and the complete followed procedure for generating the whole dataset are described in [ 43 ], and the data are available in [ 39 ].…”

Section: Methodsmentioning

confidence: 99%

“…Therefore, we present a comparative study in which we assess three marker augmenters for lower limbs across eight different poses and movements with 14 subjects, using the MediaPipe Pose Landmarker model as the pose detector [ 23 ]. We utilized the “Human tracking dataset of 3D anatomical landmarks and pose keypoints” [ 39 ], which is a collection of 3D reconstructed video keypoints associated with 3D anatomical landmarks extracted from 567 sequences of 71 participants in A-pose and performing seven movements. This dataset was designed for biomechanical applications and is freely available for the research community.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

A Deep Learning Model for Markerless Pose Estimation Based on Keypoint Augmentation: What Factors Influence Errors in Biomechanical Applications?

Ruescas-Nicolau,

Medina-Ripoll,

de Rosario

et al. 2024

Sensors

View full text Add to dashboard Cite

In biomechanics, movement is typically recorded by tracking the trajectories of anatomical landmarks previously marked using passive instrumentation, which entails several inconveniences. To overcome these disadvantages, researchers are exploring different markerless methods, such as pose estimation networks, to capture movement with equivalent accuracy to marker-based photogrammetry. However, pose estimation models usually only provide joint centers, which are incomplete data for calculating joint angles in all anatomical axes. Recently, marker augmentation models based on deep learning have emerged. These models transform pose estimation data into complete anatomical data. Building on this concept, this study presents three marker augmentation models of varying complexity that were compared to a photogrammetry system. The errors in anatomical landmark positions and the derived joint angles were calculated, and a statistical analysis of the errors was performed to identify the factors that most influence their magnitude. The proposed Transformer model improved upon the errors reported in the literature, yielding position errors of less than 1.5 cm for anatomical landmarks and 4.4 degrees for all seven movements evaluated. Anthropometric data did not influence the errors, while anatomical landmarks and movement influenced position errors, and model, rotation axis, and movement influenced joint angle errors.

show abstract

Section: Methodsmentioning

confidence: 99%

“…Further details and the complete followed procedure for generating the whole dataset are described in [ 43 ], and the data are available in [ 39 ].…”

Section: Methodsmentioning

confidence: 99%