CAJ | 학술논문

针对传统图像分割算法需要参数设置等缺点，提出了一种自动的图像分割算法，采用基于改进视觉注意机制的粗分割和结合主动轮廓与区域生长的精确分割两个过程对图像进行自动分割。实验结果表明，该方法的分割性能优于自适应阈值算法和Kmeans聚类算法，且具有较强的鲁棒性。
침대전통도상분할산법수요삼수설치등결점，제출료일충자동적도상분할산법，채용기우개진시각주의궤제적조분할화결합주동륜곽여구역생장적정학분할량개과정대도상진행자동분할。실험결과표명，해방법적분할성능우우자괄응역치산법화Kmeans취류산법，차구유교강적로봉성。
In view of the deficiency that traditional image segmentation algorithm needs to set parameters , an automatic image segmentation is proposed in this paper. It applies two process of coarse segmentation based on improved visual attention mechanism and precise segmentation combined active contour with region growing to achieve automatic image segmentation. Experiments show that the proposed algorithm outperforms adaptive threshold algorithm and Kmeans clustering algorithm in image segmentation. The robustness of the proposed algorithm is strong.