Gait analysis comparison between manual marking, 2D pose estimation algorithms, and 3D marker-based system
IntroductionRecent advances in Artificial Intelligence (AI) and Computer Vision (CV) have led to automated pose estimation algorithms using simple 2D videos. This has created the potential to perform kinematic measurements without the need for specialized, and often expensive, equipment. Even though...
| Main Authors: | , , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2023-09-01
|
| Series: | Frontiers in Rehabilitation Sciences |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/fresc.2023.1238134/full |
| _version_ | 1827823268702191616 |
|---|---|
| author | Dimitrios Menychtas Nikolaos Petrou Ioannis Kansizoglou Erasmia Giannakou Athanasios Grekidis Antonios Gasteratos Vassilios Gourgoulis Eleni Douda Ilias Smilios Maria Michalopoulou Georgios Ch. Sirakoulis Nikolaos Aggelousis |
| author_facet | Dimitrios Menychtas Nikolaos Petrou Ioannis Kansizoglou Erasmia Giannakou Athanasios Grekidis Antonios Gasteratos Vassilios Gourgoulis Eleni Douda Ilias Smilios Maria Michalopoulou Georgios Ch. Sirakoulis Nikolaos Aggelousis |
| author_sort | Dimitrios Menychtas |
| collection | DOAJ |
| description | IntroductionRecent advances in Artificial Intelligence (AI) and Computer Vision (CV) have led to automated pose estimation algorithms using simple 2D videos. This has created the potential to perform kinematic measurements without the need for specialized, and often expensive, equipment. Even though there's a growing body of literature on the development and validation of such algorithms for practical use, they haven't been adopted by health professionals. As a result, manual video annotation tools remain pretty common. Part of the reason is that the pose estimation modules can be erratic, producing errors that are difficult to rectify. Because of that, health professionals prefer the use of tried and true methods despite the time and cost savings pose estimation can offer.MethodsIn this work, the gait cycle of a sample of the elderly population on a split-belt treadmill is examined. The Openpose (OP) and Mediapipe (MP) AI pose estimation algorithms are compared to joint kinematics from a marker-based 3D motion capture system (Vicon), as well as from a video annotation tool designed for biomechanics (Kinovea). Bland-Altman (B-A) graphs and Statistical Parametric Mapping (SPM) are used to identify regions of statistically significant difference.ResultsResults showed that pose estimation can achieve motion tracking comparable to marker-based systems but struggle to identify joints that exhibit small, but crucial motion.DiscussionJoints such as the ankle, can suffer from misidentification of their anatomical landmarks. Manual tools don't have that problem, but the user will introduce a static offset across the measurements. It is proposed that an AI-powered video annotation tool that allows the user to correct errors would bring the benefits of pose estimation to professionals at a low cost. |
| first_indexed | 2024-03-12T02:09:42Z |
| format | Article |
| id | doaj.art-f546ea7f98f24a1bb5ead89111cfd8c0 |
| institution | Directory Open Access Journal |
| issn | 2673-6861 |
| language | English |
| last_indexed | 2024-03-12T02:09:42Z |
| publishDate | 2023-09-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Rehabilitation Sciences |
| spelling | doaj.art-f546ea7f98f24a1bb5ead89111cfd8c02023-09-06T15:28:40ZengFrontiers Media S.A.Frontiers in Rehabilitation Sciences2673-68612023-09-01410.3389/fresc.2023.12381341238134Gait analysis comparison between manual marking, 2D pose estimation algorithms, and 3D marker-based systemDimitrios Menychtas0Nikolaos Petrou1Ioannis Kansizoglou2Erasmia Giannakou3Athanasios Grekidis4Antonios Gasteratos5Vassilios Gourgoulis6Eleni Douda7Ilias Smilios8Maria Michalopoulou9Georgios Ch. Sirakoulis10Nikolaos Aggelousis11Biomechanics Laboratory, Department of Physical Education and Sports Science, Democritus University of Thrace, Komotini, GreeceBiomechanics Laboratory, Department of Physical Education and Sports Science, Democritus University of Thrace, Komotini, GreeceLaboratory of Robotics and Automation, Department of Production and Management Engineering, Democritus University of Thrace, Xanthi, GreeceBiomechanics Laboratory, Department of Physical Education and Sports Science, Democritus University of Thrace, Komotini, GreeceBiomechanics Laboratory, Department of Physical Education and Sports Science, Democritus University of Thrace, Komotini, GreeceLaboratory of Robotics and Automation, Department of Production and Management Engineering, Democritus University of Thrace, Xanthi, GreeceBiomechanics Laboratory, Department of Physical Education and Sports Science, Democritus University of Thrace, Komotini, GreeceBiomechanics Laboratory, Department of Physical Education and Sports Science, Democritus University of Thrace, Komotini, GreeceBiomechanics Laboratory, Department of Physical Education and Sports Science, Democritus University of Thrace, Komotini, GreeceBiomechanics Laboratory, Department of Physical Education and Sports Science, Democritus University of Thrace, Komotini, GreeceDepartment of Electrical and Computer Engineering, Democritus University of Thrace, Xanthi, GreeceBiomechanics Laboratory, Department of Physical Education and Sports Science, Democritus University of Thrace, Komotini, GreeceIntroductionRecent advances in Artificial Intelligence (AI) and Computer Vision (CV) have led to automated pose estimation algorithms using simple 2D videos. This has created the potential to perform kinematic measurements without the need for specialized, and often expensive, equipment. Even though there's a growing body of literature on the development and validation of such algorithms for practical use, they haven't been adopted by health professionals. As a result, manual video annotation tools remain pretty common. Part of the reason is that the pose estimation modules can be erratic, producing errors that are difficult to rectify. Because of that, health professionals prefer the use of tried and true methods despite the time and cost savings pose estimation can offer.MethodsIn this work, the gait cycle of a sample of the elderly population on a split-belt treadmill is examined. The Openpose (OP) and Mediapipe (MP) AI pose estimation algorithms are compared to joint kinematics from a marker-based 3D motion capture system (Vicon), as well as from a video annotation tool designed for biomechanics (Kinovea). Bland-Altman (B-A) graphs and Statistical Parametric Mapping (SPM) are used to identify regions of statistically significant difference.ResultsResults showed that pose estimation can achieve motion tracking comparable to marker-based systems but struggle to identify joints that exhibit small, but crucial motion.DiscussionJoints such as the ankle, can suffer from misidentification of their anatomical landmarks. Manual tools don't have that problem, but the user will introduce a static offset across the measurements. It is proposed that an AI-powered video annotation tool that allows the user to correct errors would bring the benefits of pose estimation to professionals at a low cost.https://www.frontiersin.org/articles/10.3389/fresc.2023.1238134/full2D pose estimationmotion analysisbiomechanicsjoint angle comparisonbiomechanics video annotation |
| spellingShingle | Dimitrios Menychtas Nikolaos Petrou Ioannis Kansizoglou Erasmia Giannakou Athanasios Grekidis Antonios Gasteratos Vassilios Gourgoulis Eleni Douda Ilias Smilios Maria Michalopoulou Georgios Ch. Sirakoulis Nikolaos Aggelousis Gait analysis comparison between manual marking, 2D pose estimation algorithms, and 3D marker-based system Frontiers in Rehabilitation Sciences 2D pose estimation motion analysis biomechanics joint angle comparison biomechanics video annotation |
| title | Gait analysis comparison between manual marking, 2D pose estimation algorithms, and 3D marker-based system |
| title_full | Gait analysis comparison between manual marking, 2D pose estimation algorithms, and 3D marker-based system |
| title_fullStr | Gait analysis comparison between manual marking, 2D pose estimation algorithms, and 3D marker-based system |
| title_full_unstemmed | Gait analysis comparison between manual marking, 2D pose estimation algorithms, and 3D marker-based system |
| title_short | Gait analysis comparison between manual marking, 2D pose estimation algorithms, and 3D marker-based system |
| title_sort | gait analysis comparison between manual marking 2d pose estimation algorithms and 3d marker based system |
| topic | 2D pose estimation motion analysis biomechanics joint angle comparison biomechanics video annotation |
| url | https://www.frontiersin.org/articles/10.3389/fresc.2023.1238134/full |
| work_keys_str_mv | AT dimitriosmenychtas gaitanalysiscomparisonbetweenmanualmarking2dposeestimationalgorithmsand3dmarkerbasedsystem AT nikolaospetrou gaitanalysiscomparisonbetweenmanualmarking2dposeestimationalgorithmsand3dmarkerbasedsystem AT ioanniskansizoglou gaitanalysiscomparisonbetweenmanualmarking2dposeestimationalgorithmsand3dmarkerbasedsystem AT erasmiagiannakou gaitanalysiscomparisonbetweenmanualmarking2dposeestimationalgorithmsand3dmarkerbasedsystem AT athanasiosgrekidis gaitanalysiscomparisonbetweenmanualmarking2dposeestimationalgorithmsand3dmarkerbasedsystem AT antoniosgasteratos gaitanalysiscomparisonbetweenmanualmarking2dposeestimationalgorithmsand3dmarkerbasedsystem AT vassiliosgourgoulis gaitanalysiscomparisonbetweenmanualmarking2dposeestimationalgorithmsand3dmarkerbasedsystem AT elenidouda gaitanalysiscomparisonbetweenmanualmarking2dposeestimationalgorithmsand3dmarkerbasedsystem AT iliassmilios gaitanalysiscomparisonbetweenmanualmarking2dposeestimationalgorithmsand3dmarkerbasedsystem AT mariamichalopoulou gaitanalysiscomparisonbetweenmanualmarking2dposeestimationalgorithmsand3dmarkerbasedsystem AT georgioschsirakoulis gaitanalysiscomparisonbetweenmanualmarking2dposeestimationalgorithmsand3dmarkerbasedsystem AT nikolaosaggelousis gaitanalysiscomparisonbetweenmanualmarking2dposeestimationalgorithmsand3dmarkerbasedsystem |