一种基于条分割及缩放的图像重定算法

An image retargeting algorithm based on strip dividing and resizing

提出了一种基于图像条分割的快速图像重定算法.根据像素梯度信息建立图像的能量图,描述图像不同区域及目标的重要性.由各图像列的累积能量,通过类间最大类内最小原理,将一幅图像分割成若干具有一定宽度和能量级的图像条.根据图像的总体缩放尺寸及不同图像条的平均能量,计算各图像条的缩减量.结合标准Scaling和Cropping方法,对图像条进行缩放操作,若边缘图像条缩减量超过一定阈值,裁剪掉该部分内容.实验结果表明,该算法能够同时保护图像的局部结构和全局视觉效果,快速生成保真的高质量重定图像.

A fast image retargeting algorithm is proposed based on strip dividing. Firstly, the image energy map is constructed by gradient information of pixels for describing the importance of the regions and objects in the image. After accumulating the energy of each column, an image is divided into several strips using the principle of minimum within-cluster difference and maximum between- cluster difference. The strip has the optimal width and individual energy level. In terms of the resizing amount of the image and the average energy of strips, the reduced amount of each strip is determined. Strips resizing is implemented through combining Scaling with Cropping. If the reduced ratio of the peripheral strip exceeds a certain threshold, then the strip will be cut. Experimental results show that the algorithm is capable of preserving both the local structures and the global visual effect of the image, producing the high-fidelity retargeted results.