摘要：Polygon approximation is a digital curve method lead to create piecewise of linear segments for digital curve. The automated determination of optimal tolerance value for a shape is a vital stage in many applications such as pattern classification and image understanding applications. This research proposes a novel model for determining the optimum tolerance value that represents the boundary shape. The method relies on the Min-# Polygon approximation criteria. It uses a recursive algorithm in combination with an interpolation model to reproduce a curve or a boundary shape from subset of original points/vertices. Then, the optimum boundary shape has been selected by applying a robust direct curve descriptor. Results indicated that the proposed method accurately detected optimum tolerance value based on minimal dominant points that reproduced the same original boundary shape. The proposed technique produces more accurate and a better identification of the optimal tolerance value of boundary shape generation based on limited points to reproduce a boundary shape leading to reduce the complexity of boundary shape determination.
