Please use this identifier to cite or link to this item:
|Title:||A THINNING ALGORITHM BY CONTOUR GENERATION|
|Authors:||Kwok, Paul C.|
|Abstract:||A fast serial algorithm for thinning binary images is presented. The boundaries of 2-D objects are chain coded and an iterative procedure is set up where the boundary points in the image are visited one after another in every iteration. A boundary point is deleted if it is not a break point or an end point. A section of the new boundary is created and contains all of the break points, end points and points exposed to the background after the previous boundary has been eroded. As the new boundary is generated it is checked for break points. At the end of an iteration, the chain code for the new boundary is in place for processing in the next iteration. The iteration terminates when all the points in the new contour are either break points or end points. The last contour generated becomes the skeleton. For thick objects, the method is 19 times faster than parallel algorithms and 6 times faster than conventional contour tracing thinning algorithms. This method is more efficient than other serial algorithms because it is faster to generate a new contour than to remove the current one and to trace a new one in the next iteration.|
|Appears in Collections:||Kwok, Paul|
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.