Gait analysis comparison between manual marking, 2D pose estimation algorithms, and 3D marker-based system.
Dimitrios Menychtas, Nikolaos Petrou, Ioannis Kansizoglou and 9 others
PMID 37744429WHAT IT FOUND
OpenPose tracked knee angles accurately but failed at the ankle.
Kinovea was consistent but slow and biased. MediaPipe was too inaccurate for gait analysis. No tool replaced the 3D marker system, though OpenPose is a promising, faster alternative for knee motion.
Key findings
01OpenPose accurately tracked knee joint angles, while MediaPipe produced large errors and misidentified joints.
02Manual annotation with Kinovea was consistent but introduced a static offset error due to difficulty identifying the hip center.
03Kinovea took about 38 hours to process 17 gait cycles, whereas OpenPose took roughly 2 minutes.
STILL TO COME
How it was doneWhat they foundWhat it means for PTsWhat it means for OTsWhat it means for SLPs
Read the rest of this summary
You get three full summaries a month, free, and we do not ask for a card. Search, the TL;DRs and your library stay unlimited either way.
What it does not show
The study used only 17 healthy, active adults, so findings may not apply to patients with mobility limitations or cognitive impairments. Only one gait cycle per person was analyzed, which may not capture natural variability in walking. Manual annotation in Kinovea was performed by a single rater, so inter-rater reliability was not assessed. The comparison relied on synchronized video from a lab setting, which may not reflect the quality of video captured in a clinic.
Declared interests
The authors declared no conflicts of interest. The work was funded by the Operational Programme “Competitiveness, Entrepreneurship and Innovation” (NSRF 2014-2020) and co-financed by Greece and the European Union.
The easy way to misread this
Do not assume that free AI pose estimation tools like OpenPose or MediaPipe are ready for clinical gait diagnosis. The study found that MediaPipe is too inaccurate for gait analysis, and OpenPose fails to reliably track the ankle, meaning these tools could miss critical lower-limb abnormalities.