Current Issue Cover
改进鱼眼变换技术的图像适应

张丽霞1, 区兆明2, 宋鸿陟1(1.华南农业大学信息学院, 广州 510642;2.广州市品高软件开发有限公司, 广州 510665)

摘 要
目的 为解决目前基于鱼眼变换技术的图像适应方法难以解决的焦点检测和多焦点冲突两大问题,提出一种基于改进鱼眼变换技术的图像适应方法。方法 提出的方法根据源图像的能量计算出图像中所有最优高能量线并组成高能量线集合,作为源图像的高能量部分,即图像的焦点区域;以能量线而不是传统的图像区域为单位进行鱼眼变换以得到目标图像。结果 改变鱼眼变换技术的变换模式并应用于图像适应中,实验结果表明,本文方法解决了基于鱼眼变换技术的图像适应方法存在的问题,通过本文算法所得到的目标图像具有较好的视觉效果,用户满意度接近4分。算法运行速度较快,将源图像(512×384)长度缩小一半的情况下仅需6 s的运算时间。结论 本文方法一方面保留了鱼眼变换图像适应方法的优势,在突出显示图像重要部分的同时,不会忽略图像的次要部分;另一方面解决了鱼眼变换图像适应方法存在的焦点检测和多焦点冲突问题。实现效果和用户主观评价结果表明,该方法是一种有效可行的图像适应方法。
关键词
Improved fisheye transformation for image retargeting

Zhang Lixia1, Ou Zhaoming2, Song Hongzhi1(1.College of Informatics, South China Agricultural University, Guangzhou 510665, China;2.Bingo Software Co., Ltd, Guangzhou 510665, China)

Abstract
Objective To solve focus detection and multi-focus problems, which are challenging issues in image retargeting based on fisheye transformation technology, a new image retargeting method that uses an improved fisheye transformation technique is proposed in this study.Method The proposed method calculates seams with high energy on the basis of source image energy. All high-energy seams are found from the high-energy part, which is the focus area of the source image. This source image is retargeted to the target image through fisheye transformation on the basis of seams instead of image regions.Result The transformation mode of the fisheye warping method is changed and applied to image retargeting. Experimental results show that our method addresses the problems of image retargeting based on fisheye transformation technology. The target images obtained through our method exhibit good visual effect. Users' subjective satisfaction score is about 4. The algorithm runs fast when the length of the target image is half of the source image (512×384). The running time is 6 s. Conclusion In this study, we propose a new method based on fisheye transformation technology for image retargeting. The method preserves the advantage of fisheye warping in retaining the focused areas of the image without discarding unfocused contents. It also uses an optimal high-energy seam finding scheme to address the problems on focus detection and multi-focus conflict. The visual effect of target images and the subjective satisfaction of users indicate that the proposed approach is a viable solution for image retargeting based on fisheye warping.
Keywords

订阅号|日报