Wound Area Segmentation Using 4-D Probability Map and Superpixel Region Growing

Chronic wounds poses a significant health burden. Gradually it's becoming a major threat for world health and economy. In developed countries, 1-2% of all population is affected by chronic wounds in their lifetime and the duration for chronic wound healing varies from several months to several years. The change in wound area over time is a highly predictive clue for wound treatment. The current clinical techniques for wound area measurement is based on visual inspection, which is potentially inaccurate and represents a tedious clinical workload for the clinician. Hence, a computerized wound measurement system is more quantitative and cost-effective that enables more accurate and effective wound management. In this paper, the proposed wound area segmentation method based on 4-D Probability map and superpixel region growing shows promising performance in terms of accuracy, stability and computational complexity. The proposed method has 79.2% accuracy, an improvement over previous method.