유전자 알고리즘과 다해상도 기법을 이용한 활성 윤곽선 모델 기반의 물체 추출

Abstract

Active contour model is so-called snake model. In this paper, we propose a object extraction based-on active contour model using a genetic algorithm for initial contour problem and effective boundary extraction. Image pyramid method used to acquire low resolution from original input image and set a initial contour. Two neighborhood parameters above set contour is selected genes and than genetic operator is applied to find object boundary. Low resolution image that is used to extract a contour enlarged step by step, compute localized active contour energy in interpolation area. Insert new contour parameter location which is minimum energy. The proposed method extracted efficiently boundary of object.