针对图像非比例缩放时出现的失真, 基于 Seam Carving算法进行了研究: 重新定义了显著能量, 以突出图像的重要内容; 优化了基于能量图的查找方法, 通过减少像素线穿越图像中重要内容的次数, 以保护图像重要内容不被扭曲; 结合双线性插值以限制抽取像素线的数目, 以避免图像内容过度抽取。实验结果表明, 改进算法能够更好的保护图像中的重要内容、提高图像缩放的质量, 并提升 Seam Carving算法的执行效率。

For image distortions during image scaling, this paper improved Seam Carving algorithm: redefined significant energy, enhancing the energy of the image to highlight the important content of the image; optimized lookup algorithm of image energy map, reducing the number of pixel lines across the image content and avoiding important image content to be distorted; combined with bilinear interpolation, reducing the number of extracted pixel line and avoiding image scaling distortion for over-abstraction of image content. The experimental results show that the improved algorithm can better protect the image content, improve the quality of image scaling, and improve the efficiency of the implementation.