CAJ | 학술논문

为快速准确地提取图像中的显著性区域，提出一种改进图割的显著性区域检测算法。采用改进的图割算法对图像进行预分割，将图像分成若干子区域，在此基础上利用区域间对比度计算各区域的显著性值，得到图像的显著图，利用迭代阈值分割算法对图像显著图进行分割处理，通过图像去噪完整有效地提取出显著性区域。实验结果表明，该算法能够准确地提取出图像中的显著性区域，与传统方法相比，使用该算法提取出的显著性区域更完整、准确，提高了效率。
위쾌속준학지제취도상중적현저성구역，제출일충개진도할적현저성구역검측산법。채용개진적도할산법대도상진행예분할，장도상분성약간자구역，재차기출상이용구역간대비도계산각구역적현저성치，득도도상적현저도，이용질대역치분할산법대도상현저도진행분할처리，통과도상거조완정유효지제취출현저성구역。실험결과표명，해산법능구준학지제취출도상중적현저성구역，여전통방법상비，사용해산법제취출적현저성구역경완정、준학，제고료효솔。
To detect the salient region quickly and accurately in the image ,the improved algorithm of salient region detection based on graph cuts was put forward .Firstly ,the improved graph cuts method was used to pre‐segment the image which was then divided into several sub areas .On this basis ,the regional contrast was used to calculate the salient value of each area which represented the salient map of image .Then the iterative threshold segmentation algorithm was used to segment the image salient map ,the salient region was extracted completely and effectively after image denoising .Experimental result shows that the im‐proved salient detection algorithm can extract the salient region accurately and completely ,it also promotes the efficiency .