A computational method for semi-automatic measurement of pressure ulcers

• Published in: Wound repair and regeneration Vol. 26; no. 4; pp. 332 - 339
• Main Authors: Silva, Rodolfo H, Machado, Alexei M C
• Format: Journal Article
• Language: English
• Published: United States 01.07.2018
• Subjects: Algorithms; Humans; Pattern Recognition, Automated; Pressure Ulcer - diagnosis; Pressure Ulcer - pathology; Reproducibility of Results; Sensitivity and Specificity; Wound Healing
• ISSN: 1067-1927; 1524-475X
• DOI: 10.1111/wrr.12650

Pressure ulcers are skin lesions caused by the excessive compression of soft tissues between bones and hard surfaces that may increase treatment risks and costs. Manual techniques to evaluate the area of the affected region include the use of adhesive labels and rulers in direct contact with the ulcer. In this article, a semi-automatic method is proposed to estimate the lesion area through computerized techniques in different color spaces using filters and morphological operators. Experiments with a set of 110 images of pressure ulcers resulted in a statistically significant increase in effectiveness, when compared with other published results, reaching 90.8% of precision, 81.6% of sensitivity, and 81.3% of accuracy.