Related Experiment Video
Updated: Oct 19, 2025
03:31
Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
Published on: December 15, 2023
695
Comprehensive Assessment of Fine-Grained Wound Images Using a Patch-Based CNN With Context-Preserving Attention
Ziyang Liu1, Emmanuel Agu1, Peder Pedersen2
1Computer Science Department, Worcester Polytechnic Institute, Worcester, MA 01609 USA.
IEEE Open Journal of Engineering in Medicine and Biology
|September 17, 2021
Summary
This study introduces an AI system using smartphone images to automatically assess chronic wounds based on the Photographic Wound Assessment Tool (PWAT) criteria, achieving over 80% accuracy for all eight wound attributes.
Area of Science:
- Medical Technology
- Artificial Intelligence
- Wound Care
Background:
- Chronic wounds impact millions, necessitating efficient remote assessment methods.
- Smartphone-based wound assessment offers a promising, accessible solution for remote monitoring.
Purpose of the Study:
- To develop and evaluate an AI system for automated, comprehensive wound grading using smartphone images.
- To assess eight key wound attributes defined by the clinically-validated Photographic Wound Assessment Tool (PWAT).
Main Methods:
- A DenseNet Convolutional Neural Network (CNN) framework with patch-based attention was employed.
- The system was trained and evaluated on a dataset of 1639 wound images.
- The model assessed diabetic ulcers, pressure ulcers, vascular ulcers, and surgical wounds.
Main Results:
- The AI model achieved classification accuracies and F1 scores exceeding 80% for all eight PWAT sub-scores.
- The system demonstrated high performance in grading various chronic wound types.
Conclusions:
- This represents the first intelligent system for autonomous, comprehensive wound grading based on PWAT criteria.
- The AI system significantly reduces the burden of manual wound grading for healthcare professionals.
