基于图像内容的缩放,目的是在对图像进行任意纵横比的缩放时保持图像的视觉主体。线裁剪是一种基于图像内容的缩放算法,在已有的线裁剪算法中,使用梯度来表示每个像素点的视觉重要性,只在对象的边界处有较高的能量值,在进行缩放时,会引起图像视觉主体的形变。本文使用图像显著图和梯度图相结合的方法来度量每个像素点的重要程度,显著图从颜色、方向和强度三个属性来计算每个像素点的视觉关注程度。实验表明,本文提出的算法在缩放中能更好地保持图像视觉主体。
Content-aware image retargeting method aim to arbitrarily change the aspect ratios of image while preserving visual prominent features.Seam carving is an image resizing algorithm based on content.To determine visual importance of pixels,existing seam carving retargeting schemes mostly rely on gradient maps.These maps show higher energy only at edges of objects and may result in deforming on salient objects.In this paper,we compute the importance of pixels through multiplying gradient value and salient value.Salient map is achieved by computing saliency of pixels using color,orientation and intensity.Experiment results show that our algorithm has a good performance in maintaining the salient objects.