Markerless Gait Deviation Analysis for Robot-Assisted Rehabilitation – Dataset and Python Pipeline

This repository contains the complete reproducible dataset and Python analysis pipeline used in the study: “Individualized Gait Deviation Analysis Using Smartphone-Based Markerless Motion Capture to Support Robot-Assisted Gait Rehabilitation.” The project includes patient-specific gait kinematic data extracted using OpenCap and processed through OpenSim 4.5 to generate joint angle trajectories. The accompanying Python pipeline provides: Gait-cycle normalization Z-score deviation mapping using a normative dataset Phase-wise deviation extraction Radar-chart visualization of gait abnormalities Automated reproducibility pipeline (master script) Processed pickle files for rapid re-analysis The dataset and code enable full reproducibility of the individualized deviation profiling for four pediatric patients (P001–P004) undergoing robot-assisted gait rehabilitation. Folder structure: /P001–P004 # Patient-specific processed kinematic data/processed_pickle # Z-score matrices, phase-wise tables (tables.pkl)/results # Saved radar charts and plots20251209_gait_robot.ipynb This Zenodo entry is intended to support transparency, reproducibility, and reuse in gait analysis and rehabilitation research.本仓库包含用于下述研究的完整可复现数据集与Python分析流程： 《基于智能手机无标记运动捕捉的个体化步态偏差分析，用于辅助机器人辅助步态康复》 本项目涵盖通过OpenCap提取、并经OpenSim 4.5处理以生成关节角轨迹的患者特异性步态运动学数据。配套的Python分析流程可实现以下功能： 1. 步态周期归一化 2. 基于标准数据集的Z分数偏差映射 3. 分阶段偏差提取 4. 步态异常雷达图可视化 5. 自动化可复现分析流程（主脚本） 6. 用于快速复分析的已处理pickle文件 本数据集与代码可完整复现4名儿科患者（P001–P004）在机器人辅助步态康复期间的个体化偏差分析结果。 文件夹结构如下： /P001–P004 # 患者特异性处理后运动学数据 /processed_pickle # Z分数矩阵、分阶段数据表（tables.pkl） /results # 已保存的雷达图与绘图文件 20251209_gait_robot.ipynb 本Zenodo存档旨在为步态分析与康复研究提供透明度、可复现性与可复用性支持。