You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Border energy attempts to find a measure of a region with high feature-to-background contrast in an attempt to discard homogeneous, less-salient regions. Regions with border energy below some threshold are discarded.
Border energy is defined as the average Sobel Edge Magnitude of all border pixels in a region.
The paper proposes a "simple adaptive threshold" not based on std-dev or variance.
The calculation for Border Energy is done. The paper mentions a 'simple adaptive threshold' as the actual decision for this region classifier. The paper is vague. Right now the decision is a dumb threshold set to an empirically chosen number
Border energy attempts to find a measure of a region with high feature-to-background contrast in an attempt to discard homogeneous, less-salient regions. Regions with border energy below some threshold are discarded.
Border energy is defined as the average Sobel Edge Magnitude of all border pixels in a region.
The paper proposes a "simple adaptive threshold" not based on std-dev or variance.
See: http://en.wikipedia.org/wiki/Sobel_operator#Formulation
for Sobel Edge detection and magnitude.
The text was updated successfully, but these errors were encountered: