University of Calgary Laboratory for Computer Vision
Software Tools
Here is a short collection of useful tools. Selected by the demand observed
on the relevant internet news groups, all programs are C source code.
Chain code:
Input is a bi-level image, supposedly containing a single object
as black pixels. The procedure returns the chain code corresponding
to the object's boundary.
Hough transform:
For a bi-level image, this function computes the Hough transform.
The resolution in Hough space can be changed to suit specific
needs.
Thin:
Our version of the famous Zhang-Suen thinning algorithm.
Signature:
Calculates the signature of an object; that is, the set of (r, theta)
values from the center of mass to the boundary of the object.
Vector templates:
For use with simple glyph-like objects. Construct a template by
vectorizing the object boundary. Result can be rescaled and
rotated when matching.
Comments and queries should be sent to parker@cpsc.ucalgary.ca.
The lab is always interested in collaboration, either industrial or
academic. We are also always interested in speaking with qualified
graduates interested in post graduate study in computer vision.