OpenCap AI於跆拳道運動表現之應用

Abstract

背景：OpenCap AI 是一種基於機器學習的動作分析系統，相較傳統 3D 運動捕捉系統，OpenCap AI無需昂貴設備，操作便利，適用於競賽與日常訓練。然而，傳統運動分析常受限於設備成本、場地限制與操作複雜性，難以彈性應用於跆拳道訓練場域。因此，本研究關注於 OpenCap AI 在跆拳道前踢動作分析中的實務應用與信效度表現。目的：本研究旨在：(1) 評估 OpenCap AI 與 Motion Analysis系統於跆拳道前踢下肢運動學參數的一致性與誤差；(2) 探討不同熱身方式對跆拳道選手前踢表現的影響。方法：招募 20 名大專跆拳道品勢選手參與三種熱身方式後執行前踢，採隨機交叉設計。以智慧型手機錄影並上傳至 OpenCap AI 系統進行運動學分析，同步以 Motion 系統進行比較。統計方法採重複量數單因子變異數分析（One-way ANOVA with Repeated Measures），並以 ICC 與 RMSE 評估兩系統之一致性與誤差表現。結果：不同熱身方式對踢擊的膝關節最大角速度與角加速度未達顯著差異（p> .05）。信效度分析顯示，OpenCap AI 與 Motion 系統間的關節角度一致性如下：膝關節： ICC = 0.991，RMSE = 4.2°（極高一致性）、髖關節： ICC = 0.937，RMSE = 8.6°（高度一致性）、踝關節： ICC = 0.860，RMSE = 15.7°（中高度一致性）討論：OpenCap AI 在膝與髖關節測量具極高一致性，顯示其可取代部分實驗室系統，並支援場域即時數據分析。儘管熱身方式未顯著影響精英選手之表現，OpenCap AI 能提供準確運動學數據，進行即時回饋與技術優化。結論：OpenCap AI 為跆拳道運動學分析提供低成本、高效率的可行方案。

Background: OpenCap AI is a machine learning-based motion analysis system offering a cost-effective and accessible alternative to traditional 3D motion capture. Its simplicity makes it suitable for Taekwondo training outside laboratory settings. This study examined the validity and practicality of OpenCap AI in analyzing front kick motions. Purpose: To (1) evaluate the agreement between OpenCap AI and a marker-based motion system in measuring lower limb kinematics during a Taekwondo front kick, and (2) assess the effects of different warm-up methods on kick performance. Methods: Twenty collegiate Taekwondo Poomsae athletes performed front kicks following three randomized warm-up conditions. Kinematic data were collected simultaneously using OpenCap AI (via smartphone video) and a marker-based system. ICC and RMSE were calculated to assess agreement; repeated measures ANOVA evaluated warm-up effects. Results: No significant differences were found in peak knee angular velocity or acceleration across warm-ups (p> .05). OpenCap AI showed excellent agreement with the motion system for the knee (ICC = 0.991, RMSE = 4.2°), good for the hip (ICC = 0.937, RMSE = 8.6°), and moderate for the ankle (ICC = 0.860, RMSE = 15.7°). Conclusion: OpenCap AI provides valid and efficient motion analysis for Taekwondo front kicks, supporting its use for real-time performance assessment in field environments.