基于双线性插值的内容感知图像缩放算法仿真

Simulation of Content-Aware Image Scaling Algorithm Based on Bilinear Interpolation

针对当前图像缩放算法存在耗时长和图像失真度高的问题,提出基于双线性插值的内容感知图像缩放算法。利用Sobel梯度算子感知图像水平与垂直方向上灰度变化,并定义图像能量函数描述图像能量分布,找到最优缝隙,为结合双线性插值算法实现图像缩放提供依据。依据图像颜色与亮度信息,根据Sobel梯度算子下的能量分布函数定义图像显著图能量函数,引入图像显著图能量增减控制阈值,突出并调节图像关键信息。融合最优能量缝隙与图像显著能量,利用双线性插值法实现图像缩放。依据逐像素、逐行生成输出图像理念,分别对横纵坐标进行线性插值,找到目标文件某点与源文件中相对应的像素点,同时对目标文件的相应图形元素进行赋值,完成图像缩放。实验结果表明,上述算法图像缩放效率高,缩放后图像失真度低。与当前算法相比,所提算法性能优越,可靠性更强。

Due to high image distortion of current algorithm,this article proposed a content-aware image scaling algorithm based on bilinear interpolation.The Sobel gradient operator was used to perceive the grayscale change in the horizontal direction and vertical direction of image.The image energy function was defined to describe the image energy distribution,so as to find the optimal gap,which provided a basis for the image scaling with bilinear interpolation algorithm.Based on the image color and brightness information,the energy function of image saliency graph was defined by the energy distribution function under Sobel gradient operator.The energy of image saliency map was introduced to increase and decrease control threshold,so as to highlight and adjust the key information of image.By integrating the optimal energy gap with the salient energy of image,the bilinear interpolation was used to achieve the image scaling.According to the concept of pixel-by-pixel and line-by-line generation output,horizontal and vertical coordinates were linearly interpolated,so that the pixel point of a point in target file corresponding to the source file was found.Finally,the values were assigned to the corresponding graphic element of target file.Thus,we completed the image scaling.Simulation results verify that the proposed algorithm has high image scaling efficiency and low image distortion after scaling.Compared with the current algorithm,the proposed algorithm has superior performance and stronger reliability.