Related Experiment Video
Updated: Jul 10, 2026
06:36
Murine Excisional Wound Healing Model and Histological Morphometric Wound Analysis
Published on: August 21, 2020
Supervised tissue classification from color images for a complete wound assessment tool
Hazem Wannous1, Sylvie Treuillet, Yves Lucas
1University of Orleans, ENSI of Bourges, Vision and Robotics Lab, Bd Lahitolle 18000 Bourges, France. hazem.wannous@ensi-bourges.fr
Summary
This study developed a new method for wound tissue color classification using unsupervised segmentation before classification. This approach improves wound assessment accuracy for granulation and slough, aiding in better wound management.
Area of Science:
- Medical Imaging
- Computer Vision
- Biomedical Engineering
Background:
- Accurate wound assessment is crucial for effective treatment.
- Existing methods for wound tissue analysis have limitations in color classification.
- A 3D and color wound assessment tool is being developed within the ESCALE project.
Purpose of the Study:
- To develop and validate a robust color classification method for wound tissues.
- To improve the accuracy of wound tissue labeling by incorporating spatial information.
- To lay the groundwork for a combined 3D shape and color wound analysis tool.
Main Methods:
- An original approach using unsupervised segmentation prior to classification was employed.
- A ground truth was established using clinician-labeled images.
- Support Vector Machine (SVM) region classifiers were trained using color and texture descriptors.
- Unsupervised color region segmentation was applied to test images for classification.
Main Results:
- The SVM region classifier achieved an 88% success overlap score.
- Segmentation-driven classification showed comparable performance to clinician labeling.
- Accuracies of approximately 75% for granulation and 60% for slough were achieved.
Conclusions:
- Unsupervised segmentation prior to classification enhances the robustness of wound tissue labeling.
- The developed method shows promise for accurate color classification of wound tissues.
- This work is a significant step towards a comprehensive 3D and color wound assessment tool.
