Smart Insole-Based Plantar Pressure Analysis for Healthy and Diabetic Feet Classification: Statistical vs. Machine Learning Approaches

Technologies, Journal Year: 2024, Volume and Issue: 12(11), P. 231 - 231

Published: Nov. 19, 2024

Diabetes is a significant global health issue impacting millions. Approximately 26 million diabetics experience foot ulcers, with 20% ending up amputations, resulting in high morbidity, mortality, and costs. Plantar pressure screening shows potential for early detection of Diabetic Foot Ulcers (DFUs). Although ulcers often occur due to excessive on the soles during dynamic activities, most studies focus static measurements. This study’s primary objective apply wireless plantar sensor-embedded insoles classify detect diabetic feet from healthy ones based pressure. The secondary compare statistical-based Machine Learning (ML) classification methods. Data 150 subjects were collected walking, revealing that have higher than feet, which consistent prior research. Adaptive Boosting (AdaBoost) ML model achieved highest accuracy 0.85, outperforming statistical method, had an 0.67. These findings suggest models, combined insoles, can effectively using features. Future research will these various stages neuropathy, aiming prediction home settings.

Language: Английский

Enhanced Diabetes Detection from Foot Plantar Thermographs Using an Attention-Infused InceptionV3 Residual Block

International Journal of Electronics and Communication Engineering, Journal Year: 2024, Volume and Issue: 11(11), P. 257 - 271

Published: Nov. 30, 2024

chronic illness, Diabetes Mellitus (DM), occurs due to the inability of pancreas produce insulin or utilize it produces effectively. People with diabetes have increased risks developing various life-threatening conditions, resulting in reduced quality life and mortality. causes long-term impairment degradation many body parts. Early intervention treatment can prevent extreme outcomes such as amputation. Thermography is a non-invasive technique commonly used detect variations temperature distribution foot region. So, this study, hybrid Deep Learning (DL) model incorporating pretrained inception V3 custom layers attention residual blocks proposed from plantar thermographic images efficiently. dataset are utilized study preprocessing data augmentation techniques. The exhibits superior performance when compared state-of-the-art methods 95.71% accuracy, 97.85% precision, 93.83% recall, 95.80 % F1score. In addition standard evaluation metrics, DL models measured Cohen's kappa Area under Curve (AUC). indicate model's potential real-time clinical application, more effective diabetic detection management.

Language: Английский