A Comprehensive Review of Machine Learning and Deep Learning Approaches for Early Detection of Diabetes Mellitus Using Gait Analysis and Plantar Pressure Data

Diabetes Mellitus (DM) represents a major global health challenge, with early detection being crucial to prevent severe complications such as diabetic foot ulcers (DFUs) and neuropathy. Traditional diagnostic methods often rely on invasive tests, but recent advancements in non-invasive techniques, particularly gait analysis and plantar pressure assessment, offer promising alternatives. This systematic review examines the application of machine learning (ML) and deep learning (DL) models in processing biomechanical data from gait and plantar pressure sensors for early DM detection. We categorize data acquisition methods, evaluate key ML/DL architectures, analyze performance metrics, and identify gaps in current research. Drawing from ۳۰ selected studies spanning ۲۰۱۶ to ۲۰۲۵, the review highlights how these approaches achieve accuracies up to ۹۸% while addressing challenges like data scarcity and model generalization. Recommendations for future work emphasize hybrid models and wearable integration for real-time monitoring.

کلیدواژه ها:

Diabetes Mellitus ، Early Detection ، Gait Analysis ، Plantar Pressure ، Machine Learning ، Deep Learning ، Biomechanical Sensors ، Systematic Review