Smart Wound Monitoring and Healing Assessment System With Deep Learning Methods - A Systematic Review

Abstract

This review describes a Deep Learning Smart Wound Monitoring (DL-SWM) system, focusing on the healing process of both acute and chronic skin wounds. Recognizing the impact of various factors such as environment, patient characteristics, and wound features on the healing timeline, the review emphasizes the need for more efficient wound monitoring methods. The proposed DL-SWM integrates biosensors, a microcontroller, and a fuzzy inference system to assess critical wound indicators, primarily focusing on hydration levels. The hardware design incorporates an Arduino-based biometric sensor device, while the fuzzy inference system predicts the impact of biomarkers on wound hydration. The review study also explores the segmentation of wounds using a Convolutional Neural Network (CNN) called MobileNetV2, providing detailed insights into the wound healing stages. In the literature review, various advancements in wound monitoring technologies, such as hydrogels, clinical decision-making systems, and wearable biological sensors, are discussed. The proposed DL-SWM is compared with existing methods through simulation analysis, demonstrating superior efficiency, accuracy, and lower error rates. The study concludes with the potential prospects of DL-SWM in revolutionizing wound monitoring and treatment, offering a more convenient and effective approach for healthcare practitioners and improving patient outcomes.