Skip to main content
. 2021 Jan 19;12:458. doi: 10.1038/s41467-020-20730-y

Fig. 1. Visibility graphs as a descriptor of shape.

Fig. 1

a Contour of an illustrative shape. Nodes (blue) are placed equidistantly along the contour and used to create a visibility graph. Two nodes are connected in the visibility graph if they can “see each other”, i.e., the segment connecting them does not cross or align with the contour (teal edge); otherwise, the nodes cannot see each other (orange edge). The edges between nodes can be weighted according to their edge length (Euclidean distance dE). b Visualization of the visibility graph for the cell in (a). c Heatmap of the weighted and scaled adjacency matrix of the visibility graph (a), where edges with a weight of zero indicate the absence of said edge, thus indicating “non-visibility” between the nodes. The weighted visibility graph is used for local feature extraction, while the unweighted graph is used for global shape description and shape comparison. Source data are provided as a Source Data file.