Lower Limb Gait Analysis Using Deep Learning

DOI

10.2991/978-94-6463-858-5_249How to use a DOI?

Keywords

Gait Analysis; Deep Learning; CNN-LSTM; EMG Signals; Neural Networks; Lower Limb Disorders

Abstract

Understanding human gait patterns is essential for medical diagnostics, rehabilitation, and prosthetic development. This study presents an advanced deep learning approach that combines Convolutional Neural Networks (CNNs), Long Short-Term Memory (LSTM), and Gated Recurrent Units (GRUs) to analyze lower limb movement using surface Electromyography (sEMG) signals. The dataset comprises multiple sEMG recordings from lower limb muscles alongside knee flexion-extension measurements, which were preprocessed through normalization, segmentation, and data partitioning. The hybrid model utilizes CNN layers for spatial feature extraction, while bidirectional LSTM and GRU layers capture temporal dependencies in gait signals, leading to enhanced prediction accuracy. The preprocessing phase involved scaling the sEMG data using minimax normalization and segmenting it into fixed-length sequences to maintain consistency. The hybrid deep learning architecture was carefully designed to process these sequences, leveraging CNN for spatial pattern recognition and the LSTM-GRU combination for learning sequential dependencies. Model training was performed using the AdamW optimizer with Mean Squared Error (MSE) as the loss function, incorporating early stopping to prevent overfitting and ensure robust generalization. Extensive experiments were conducted to evaluate the model’s performance using an 80–20 train-test split. Metrics such as the R² score and Mean Absolute Error (MAE) were employed to measure predictive accuracy. The results demonstrate that the proposed hybrid model effectively captures gait characteristics, achieving high accuracy in knee flexion-extension predictions. A comparison between actual and predicted values confirms its suitability for gait analysis applications. This study underscores the potential of deep learning in biomechanical research and suggests future advancements, including real- time gait monitoring and adaptive rehabilitation solutions.

Copyright

© 2025 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Tanaya Kanungo
AU  - J. Venekha
AU  - V. Karpagalakshmi
AU  - S. Sugumaran
PY  - 2025
DA  - 2025/11/04
TI  - Lower Limb Gait Analysis Using Deep Learning
BT  - Proceedings of International Conference on Computer Science and Communication Engineering (ICCSCE 2025)
PB  - Atlantis Press
SP  - 2971
EP  - 2982
SN  - 2352-538X
UR  - https://doi.org/10.2991/978-94-6463-858-5_249
DO  - 10.2991/978-94-6463-858-5_249
ID  - Kanungo2025
ER  -