GaitRec-Net: A Deep Neural Network for Gait Disorder Detection Using Ground Reaction Force.

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.