“…Edge detection segmentation regionalizes an image by locating feature boundary elements, while region-growing algorithms make use of seed pixels that are grown into objects, based on a region's homogeneity (Fortin et al, 2000;Munoz et al, 2003;Blaschke et al, 2014). Many factors affect the quality of segmentation results (Moller et al, 2007), and much research has been done on improving image object delineation by considering different segmentation algorithms (Meinel & Neubert, 2004), tuning input parameters (Achanccaray et al, 2014;Achanccaray et al, 2015) and studying the effect that the spatial and spectral characteristics of the objects of interest have on the output (Lucieer & Stein, 2002;Burnett & Blaschke, 2003;Tian & Chen, 2007;Blaschke, 2010;Petropoulos et al, 2012). Effective validation techniques that quantify a segmentation result's ability to delineate objects of interest remain elusive (Lucieer & Stein, 2002;Benz et al, 2004;Clinton et al, 2010).…”