提出一种利用视觉显著性对图像进行分割的方法。首先提取图像的底层视觉特征，从局部显著性、全局显著性和稀少性3个方面计算各特征图像中各像素的视觉显著性，得到各特征显著图；对各特征显著图进行综合，生

成最终的综合显著图。然后对综合显著图进行阈值分割，得到二值图像，将二值图像与原始图像叠加，将前景和背景分离，得到图像分割结果-It presents a significant advantage of visual image segmentation method. First, extract the underlying visual features of images,　the local significant global significance and rarity three terms calculated for each feature image each pixel visual significance to give each feature saliency map　for each feature saliency map integrated to generate the final integrated saliency map. Then the integrated saliency map threshold segmentation, binary image, binary image with the original image overlay, foreground and background separation, image segmentation results obtained