Related Experiment Video
Updated: Jul 19, 2025
11:35
Dual-mode Imaging of Cutaneous Tissue Oxygenation and Vascular Function
Published on: December 8, 2010
16.6K
Artificial Intelligence Optical Biopsy for Evaluating the Functional State of Wounds
Joe Teague1, Damien Socia1, Gary An1
1Department of Surgery, University of Vermont, Burlington, Vermont.
The Journal of Surgical Research
|August 10, 2023
Summary
This study introduces a novel machine learning approach using Siamese neural networks (SNNs) to predict wound healing status from digital images, bypassing the need for tissue biopsies.
Area of Science:
- Biomedical Engineering
- Computational Biology
- Wound Healing Research
Background:
- Assessing active wound functional status requires invasive tissue biopsies.
- Understanding cellular and molecular drivers of wound healing is clinically challenging.
Purpose of the Study:
- To develop a non-invasive method for predicting wound functional status using digital imaging.
- To evaluate the efficacy of a Siamese neural network (SNN) architecture in wound analysis.
Main Methods:
- Utilized a canine model of volumetric muscle loss (VML) with digital wound images and tissue biopsies.
- Performed RNA sequencing on biopsies to generate gene expression and functional profiles.
- Trained an SNN to regress functional profiles from wound image segments.
Main Results:
- The SNN accurately predicted functional profiles with errors between 5% and 30%.
- Functions associated with early wound healing stages were predicted most effectively.
Conclusions:
- This pilot study demonstrates the potential of machine learning regression on medical images for wound assessment.
- Regressing functional profiles offers a robust alternative to gene-specific analysis, providing deeper mechanistic insights.
