the related concepts and applications of edge detection in ecology are both useful
and interesting, they also represent a relatively well-defined body of knowledge
(Forman, 1995). However, the notion of detecting blobs and their association with
multiscale object-based frameworks (Hay et al., 2002a; Burnett and Blaschke,
2003; Syed et al., 2011) and the hierarchical patch dynamics paradigm (Wu,
1999) represent a more recent body of multiscale research where exciting
opportunities and new discoveries still exist. It is this avenue of research that
we will report on here.
To delineate dominant features within the SS stack, we apply blob–feature
detection. Its primary objective is to link structures (i.e., gray-level blobs) at different
scales in SS, to higher order objects called SS blobs and to extract significant features
based on their appearance and persistence through scales. When blobs are evaluated
as a volumetric structure within a stack, it becomes apparent that some structures
visually persist through scale, while others disappear (Figure 8.6). Consequently, an
important premise of SS is that bloblike structures which “persist” in SS are likely
candidates to correspond to significant structures in the image and thus in the realworld scene.
FIGURE 8.6 Gray-level stack with opacity filters applied to illustrate persistence of blob
structures through scale with original image on bottom to provide context.
152
VISUALIZING SCALE-DOMAIN MANIFOLDS
and interesting, they also represent a relatively well-defined body of knowledge
(Forman, 1995). However, the notion of detecting blobs and their association with
multiscale object-based frameworks (Hay et al., 2002a; Burnett and Blaschke,
2003; Syed et al., 2011) and the hierarchical patch dynamics paradigm (Wu,
1999) represent a more recent body of multiscale research where exciting
opportunities and new discoveries still exist. It is this avenue of research that
we will report on here.
To delineate dominant features within the SS stack, we apply blob–feature
detection. Its primary objective is to link structures (i.e., gray-level blobs) at different
scales in SS, to higher order objects called SS blobs and to extract significant features
based on their appearance and persistence through scales. When blobs are evaluated
as a volumetric structure within a stack, it becomes apparent that some structures
visually persist through scale, while others disappear (Figure 8.6). Consequently, an
important premise of SS is that bloblike structures which “persist” in SS are likely
candidates to correspond to significant structures in the image and thus in the realworld scene.
FIGURE 8.6 Gray-level stack with opacity filters applied to illustrate persistence of blob
structures through scale with original image on bottom to provide context.
152
VISUALIZING SCALE-DOMAIN MANIFOLDS