Analysis and Implementation of Multimodal Gait Recognition Using Logistic Regression for Neurodegenerative Disease

In this work, a multimodal gait recognition system is developed for the classification of neurodegenerative diseases such as Parkinson’s disease, Huntington’s disease, and ALS by using NDD and PD gait datasets. The system applies spatial feature extraction, deep learning, and machine learning techniques to achieve accurate classification of gait patterns. The results show that the deep learning model provides the highest accuracy, while machine learning models such as SVM, Random Forest, and Logistic Regression also give stable performance. Logistic Regression is selected for FPGA implementation, and the trained model is successfully deployed using a custom AXI-based IP on the PYNQ-Z2 board for real-time processing. Overall, the proposed system provides an efficient and reliable solution for gait-based disease detection and can be useful for clinical applications, and in future work, performance can be improved by using larger datasets and advanced models.

[1] A. Zhao et al., “Multimodal Gait Recognition for Neurodegenerative Diseases,” IEEE Trans. Cybernetics, 2022  [2] L. Shi et al., “Deep Learning Network for Gait Recognition Using Multimodal Sensors,” Sensors, 2023. [3] J. Li et al., “CNN-BiLSTM Based Multimodal Gait Recognition,” Sensors, 2023. [4] H. Wang et al., “Gait Recognition for Parkinson’s Disease,” Scientific Reports, 2025. [5] K. Açıcı et al., “Random Forest for Parkinson Detection Using Gait,” Springer, 2017. [6] A. AlAsadi, “Gait Recognition Using SVM and Neural Networks,” 2014. [7] J. Dong et al., “Deep Multimodal Gait Recognition,” IEEE, 2021. [8] S. Yu et al., “GaitBase: A Robust Gait Recognition Model,” 2022. [9] S. Zou et al., “Multi-Stage Adaptive Feature Fusion for Gait Recognition,” 2023. [10] M. Alnaasan et al., “Explainable Multimodal Gait Recognition Using RGB-D,” 2025