This study presented a deep learning based markerless motion capture workflow and evaluated performance against marker-based motion capture during overground running. Multi-view high speed (200 Hz) image data were collected concurrently with marker-based motion capture (ground-truth data) permitting a direct comparison between methods. Lower limb kinematic data for six participants demonstrated high levels of agreement for lower limb joint angles with average RMSE ranging between 2.5° - 4.4° for hip sagittal and frontal plane motion, and 4.2° - 5.2° for knee and ankle motion. These differences generally fall within the known uncertainties of marker-based motion capture, suggesting that our markerless approach could be used for appropriate biomechanics applications. While there is a need for high quality open-access datasets to further facilitate performance improvements, markerless motion capture technology continues to improve; presenting exciting opportunities for biomechanics researchers and practitioners to capture large amounts of high quality, ecologically valid data both in and out of the laboratory setting.qual
"DEVELOPMENT AND EVALUATION OF A DEEP LEARNING BASED MARKERLESS MOTION CAPTURE SYSTEM,"
ISBS Proceedings Archive: Vol. 39:
1, Article 32.
Available at: https://commons.nmu.edu/isbs/vol39/iss1/32