Gait Analysis Software

Overview

This project aims to facilitate gait analysis for individuals with lower limb amputations. The primary objective is to detect gait deviations and provide solutions for correction. The software is designed with a user-friendly interface, making it accessible for healthcare professionals involved in the analysis of biomechanics during walking.

Features

• Real-time visualization of hip, knee, and ankle angles for both left and right legs.
• Gait deviation analysis tailored for lower limb amputees.
• User-friendly interface for intuitive navigation and understanding of gait data.

Getting Started

Python

This code was developed and tested using Python 3.8.10. While it may work with other Python 3 versions, it is recommended to use Python 3.8.10 for optimal compatibility.

Check your Python version by running the script:

python --version

Object Detection

Object Detection Model

The object detection model used in this project was initially trained with YOLOv3. Please note that since the time of the initial training, more efficient versions of the YOLO algorithm have been developed, which may offer improved performance.

Training Object and Anatomical Points

The object used for training the object detection model is a specific yellow marker. During data collection, it is essential to place this marker at specific anatomical points to ensure accurate detection. The marker should be visible during the majority of the gait sequence.

3D Design and Printing:

The training objects were 3D designed using Fusion 360, and the final models were 3D printed using yellow TPU (Thermoplastic Polyurethane) material. The choice of yellow is intentional, as it provides high contrast with the majority of environments and clothing. This color selection enhances the robustness of the detection model, making the yellow marker easily distinguishable.

Anatomical Points:

1. Acromium

2. Femur

3. Femur Lateral Epicondyle

4. Tibial Lateral Malleolus

5. Metatarsal

Placement Guidelines:

• Carefully position the yellow marker at each anatomical point during data collection.
• Ensure that the marker remains visible for the majority of the gait sequence to facilitate accurate object detection.

These specific placements are crucial for training and evaluating the model's performance. Adjust the camera angle and placement as needed to achieve optimal visibility of the yellow marker at the specified anatomical points.

Clone the Github repository

git clone https://github.com/COU7INHO/Gait_Analysis_Software.git

Install the requirements

To install the requirements run the command:

pip install -r requirements.txt

Database setup

As the database is running locally, when you run the script for the first time, the database will be empty. To test the software effectively, make sure to create some data, patients, in the database.

Warning

Due to computing limitations, the video analysis may run at a very low FPS rate. Please be aware of potential performance issues and consider running the software on a system with sufficient computational resources.