基于 Snake 和外观模板的组合式图像对象分割

摘要

For image feature extraction or abstract degree of the previous image object segmentation model method is not enough,and be abandoned too much image information,this paper proposed a method that combined image segmentation based on Snake model and the appearance of the template model method.The method made the result of the appearance of template image object segmentation algorithm after pretreatment as initial contour for precise processing of the Snake model of image segmentation algorithm.The effectiveness of the proposed algorithm in ascension object segmentation accuracy reaches more than 96%.%针对目前图像对象分割方法对图像的特征提取和抽象度不够，导致分割精度低的问题，提出了基于Snake 模型和外观模板的组合式图像对象分割研究方法。该方法将改进的外观模板图像对象分割算法得到的分割结果经过预处理作为改进的 Snake 模型的图像分割算法的输入初始轮廓进行精确处理。该算法在提升对象分割精度的有效性上达到了96％以上。