CornerPoint

Skeleton corner detection from a binary image based on a 3x3 region.

 

function call:

[out_img1]=vsg('CornerPoint',in_img1);

 

Arguments:

in_img1 input image, 3 channel RGB, 1 channel greyscale or binary image, or DICOM image.

 

Description:

The CornerPoint function uses 3x3 masks of varying orientation to detect corner points in a binary image. It detects 3×3 corner points and replaces them by a single white pixel if the illustrated patterns occur (Ref). Otherwise, the output pixel (shaded) is set to black. out_img1 - An image which has dimensions the same as the input image. The output image displays the binary corner points.

 

Example:

img1 = openimage('A.bmp'); %color image

[out_img1]=vsg('Threshold',img1,100);

figure;image(uint8(out_img1));

[out_img1]=vsg('CornerPoint',out_img1);

h=figure;image(uint8(out_img1));

Notes:

For non-binary images or colour-planes/slices, corner points must have exactly 0 and 255 values within the corner point detection kernel for that point to be flagged as a corner.

 

Ref: Paul F. Whelan and Derek Molloy, “Machine Vision Algorithm in Java – Techniques and implementation”, Springer, 2000.
Home       Functions       Back