Back to results
US 10,096,097 B2Granted
Content-aware bidirectional image edge highlighting
AssigneeThe United States of America as represented by the Secretary of the Navy
InventorsGregory Knowles Fleizach, Brian Matthew Sperlongano, Kevin Donald Casey
Life of this patent
Estimating the term…
Abstract
A method for highlighting edges of interest in an image comprising the following steps: calculating edge intensity of edges in the image using gradient operators; applying a transformation to emphasize mid-level edges and deemphasize strong edges; and if a given transformed edge is greater than a threshold value, highlighting the corresponding edge in an edge-highlighted image.
Claims9
A method for highlighting edges of interest in an image comprising the following steps:
- calculating edge intensity of the image using horizontal and vertical gradient operators, wherein the image comprises rows and columns of pixels;
- combining results from the horizontal and vertical gradient operators to produce a single image, which contains the relative intensity strengths (edge_strength) of all the edges in the image;
- transforming the edge_strength image according to the following equation:
- if
-
- (
- edge_strength
-
- [
- n
- ,
- m
- ]
- >
- max
-
- (
- edge_strength
- )
- /
- 2
- )
- {
-
- edge_strength
-
- [
- n
- ,
- m
- ]
- =
- max
-
- (
- edge_strength
- )
- -
- edge_strength
-
- [
- n
- ,
- m
- ]
- ;
-
-
- }
- where edge_strength[n,m] is the edge strength value of a pixel at row n, column m, and max(edge_strength) is a maximum value of edge_strength over the entire image such that strong intensity edges are deemphasized;
- normalizing results of the transformation to put the transformed edge strength values into the range [0,1];
- and
- bidirectionally highlighting a given edge in the original image if the normalized transformed edge is greater than a threshold value to create an edge-highlighted image.
50 more elements
Dependencies are inferred from the claim text.