Wound Healing Modeling Using Partial Differential Equation and Deep Learning

TL;DR

This paper presents a novel approach combining PDE modeling with deep learning to analyze and predict wound healing processes from images, aiming to improve personalized treatment strategies.

Contribution

It introduces the first integration of numerical PDE solutions with deep learning for automated wound healing analysis from images.

Findings

The combined PDE and deep learning model achieves reasonable accuracy in wound healing prediction.

The approach enables automated segmentation and analysis of wound images.

This method offers a new direction for personalized medicine in wound care.

Abstract

The process of wound healing has been an active area of research around the world. The problem is the wounds of different patients heal differently. For example, patients with a background of diabetes may have difficulties in healing [1]. By clearly understanding this process, we can determine the type and quantity of medicine to give to patients with varying types of wounds. In this research, we use a variation of the Alternating Direction Implicit method to solve a partial differential equation that models part of the wound healing process. Wound images are used as the dataset that we analyze. To segment the image's wound, we implement deep learning-based models. We show that the combination of a variant of the Alternating Direction Implicit method and Deep Learning provides a reasonably accurate model for the process of wound healing. To the best of our knowledge, this is the first attempt to combine both numerical PDE and deep learning techniques in an automated system to capture the long-term behavior of wound healing.