Naoto Ienaga, Kazuki Sekine. "Comparison of deep learning-based three-dimensional human pose estimation methods with motion capture for gesture research." PLoS ONE (2026). https://doi.org/10.1371/journal.pone.0347288