Figure 1. Overview of Methodology for Evaluation of Artificial Intelligence (AI)–Based Digital Wound Assessment Tools.

A, Example images of wound area and granulation tissue tracings by humans and AI for wounds of diverse types, shapes, and sizes. B, Illustration of quantitative method for comparing wound area and granulation tissue tracings between humans and between humans and AI. One human tracing was selected as a reference trace (R), a second tracing (other human or AI) was selected as the test trace (T), and the overlapping area (O) was determined. The error measures (false-negative area [FNA], false positive area [FPA], and absolute relative error [ARE]) between the reference and test tracings were then calculated.