The output of ss_edges.m (and ss_edge_worm.m) is a matlab file named like edges_[filestem].mat.  The variables stored in this file contain a lists of edge points, edge segments, nodes (intersections of edge segments), relationships between points/segments/nodes, edge point properties, and maps of edge point properties.

The program ss_display_edges.m is provided as an example of how the edge data can be "cleaned" -- in this case simple thresholds of simple salience measures -- and/or displayed. ss_display_edges.m is called at the end of ss_edges.m but can be run separately. It displays a series of matlab figures containing a graphical representation of the edge segment data.  An example of (most) possible figure window contents is given below with a brief description. See the header of ss_display_edges.m for more information on selecting display and threshold parameters.
 
 
 

Resampled image (see rat MRI example).  The size of the display relative to the resampled image size, can be selected with the third parameter of ss_display_edges.m.  Left click to zoom (on any output window).

 
 

Gradient map (magnitude of maximal gradient in scale) at all edge point locations.  Very low gradients are too dim to display.

 
 
 

Scale map at all edge point locations.  The medium gray values are at the lowest scales, and brighter values at large scales.  Most variation in scale can't be distinguished, but it allows quick identification of large scale segments due to "staircase" edges.  This image has no extended soft (wide, or exclusively large scale) edges.

 
 
 

Vector edges overlaid onto slightly dimmed resampled image.  Edge segments are yellow, with the "strength" (gradient magnitude) coded by contrast with the background. Note that threshold values have suppressed low integrated gradient, short, and larger scale edge segments.

 
 
 

Scatter plot of edge segment salience measures. Red lines indicate thresholds (see parameters to ss_display_edges.m).  Average scale threshold is not indicated.