Adaptive prototype generating technique for improving performance of a p-Snake

Abstract p-Snake is an energy minimizing algorithm that applies an additional prototype energy to the existing ActiveContour Model and is used to extract the contour line in the area where the edge information is unclear. In this papersuggested the creation of a prototype energy field that applies a variable prototype expressed as a combination ofcircle and straight line primitives, and a fudge function, to improve p-Snake’s contour extraction performance. The prototype was defined based on the parts codes entered and the appropriate initial contour was extracted in each primitive zones acquired from the pre-processing process. Then, the primitives variably adjusted to create the prototype and the contour probability based on the distance to the prototype was calculated through the fuzzy functionto create the prototype energy field. This was applied to p-Snake to extract the contour from 100 images acquiredfrom various small parts and compared its similarity with the prototype to find that p-Snake made with the adaptive prototype was about 4.6% more precise than the existing Snake method.