Development and Analysis of Dermal Wound Image Processing Techniques Using Chan-Vese and Edge Active Contour Methods Public Deposited

Downloadable Content

Download PDF
Last Modified
  • March 21, 2019
Creator
  • Mueller, Peter
    • Affiliation: School of Medicine, UNC/NCSU Joint Department of Biomedical Engineering
Abstract
  • Present methods for evaluation of burn wounds rely heavily on qualitative and Total Body Surface Area (TBSA) estimations. Herein, a digital method for calculating the surface area of burn wounds is proposed as a useful tool for monitoring and measuring changes in burns. This study tested two segmentation methods: a statistical analysis technique, and an active contours technique using edges and Chan-Vese. All methods were tested on images of burns taken from a DSLR camera, and Microsoft Kinect V2 and compared to digitally drawn traces of the wounds. Using Dice’s Coefficient as a measure for agreement between masks, the DSLR images resulted in agreeable segmentations (D=0.939 for edge, D=0.9362 for Chan-Vese), while images taken with the Kinect did not meet the threshold for agreeability (D=0.815 for edge, D=0.819 for Chan-Vese). This testing shows the active contours method is the plausible method for characterizing high-resolution color burn wounds.
Date of publication
Keyword
DOI
Resource type
Rights statement
  • In Copyright
Advisor
  • Hubbard, Devin
  • Gomez, Shawn
  • Walker, Dwight
  • Pinton, Gianmarco
Degree
  • Master of Science
Degree granting institution
  • University of North Carolina at Chapel Hill Graduate School
Graduation year
  • 2017
Language
Parents:

This work has no parents.

Items