GaitRec-Net: A Deep Neural Network for Gait Disorder Detection Using Ground Reaction Force
Chandrasen Pandey1, Diptendu Sinha Roy1, Ramesh Chandra Poonia2
1National Institute of Technology, Meghalaya, India.
PPAR Research
|September 1, 2022
Summary
This study introduces an AI-driven method using ground reaction force (GRF) to automatically detect gait disorders. The proposed deep learning model, GaitRec-Net, achieves high accuracy in classifying abnormal gait patterns.
Area of Science:
- Biomechanics
- Artificial Intelligence
- Medical Diagnostics
Background:
- Gait irregularities are key indicators of neurological and musculoskeletal disorders.
- Traditional gait analysis methods (video, pressure mats) are often complex and resource-intensive.
- Accurate gait assessment is crucial for early diagnosis and effective treatment planning.
Purpose of the Study:
- To develop and evaluate an artificial intelligence (AI)-based framework for classifying gait disorders using ground reaction force (GRF) patterns.
- To compare the performance of machine learning (ML) and deep learning (DL) models for gait disorder classification.
- To introduce a novel deep learning architecture, GaitRec-Net, for enhanced gait analysis.
Main Methods:
- Utilized a large-scale dataset of ground reaction force (GRF) measurements from healthy individuals and patients with gait disorders.
- Implemented and compared various machine learning (ML) classifiers.
- Developed and applied a novel deep learning architecture, GaitRec-Net, for GRF pattern classification.
- Employed a fivefold cross-validation approach for rigorous evaluation of all models.
Main Results:
- The proposed deep learning model, GaitRec-Net, demonstrated superior performance compared to traditional ML classifiers.
- GaitRec-Net achieved high accuracy in classifying healthy controls and individuals with gait disorders.
- The deep learning approach proved more effective for feature extraction from GRF data, leading to improved classification outcomes.
Conclusions:
- The AI-based framework, particularly the GaitRec-Net model, offers a promising approach for the automatic and accurate categorization of abnormal gait patterns.
- This technology has the potential to significantly aid in the early detection and management of gait-related disabilities.
- GRF analysis combined with deep learning presents a powerful tool for clinical gait assessment.
