信息与控制2017，Vol.46Issue(4)：422-427,6.DOI:10.13976/j.cnki.xk.2017.0422
高阶变分正则化的图像放大算法
Image Zooming Method Based on High-order Variation
摘要
Abstract
To address edge blur and the blocky effects problem in the process of zooming in on an image on the basis of second-order differential equation, we propose a new image zooming method according to characteristic of denoising and edge protection for a high-order variation differential equation.First, in accordance with the Chambolle image zooming model, we regard the original low-resolution image as the subspace projection of a zoomed image to its own space.Second, we construct a new image zooming model by introducing a Laplace variation regular term that contains edge detection function.Finally, we introduce an efficient and fast split Bregman algorithm based on an analysis of the properties of the new model.The final zooming image is obtained.Experiments show that the proposed algorithm based on high-order variation can avoid blocky effects, thereby significantly enhancing the texture and the edges of a zoomed image and significantly improving visual effects as a result.关键词
图像放大/高阶微分方程/变分/边缘检测函数/分裂Bregman算法Key words
image zooming/high-order differential equation/variation/edge detection function/split Bregman algorithm分类
信息技术与安全科学引用本文复制引用
刘晨华,颜兵..高阶变分正则化的图像放大算法[J].信息与控制,2017,46(4):422-427,6.基金项目
国家青年自然科学基金资助项目(11501401) （11501401）
太原科技大学校青年科技研究基金资助项目(20133011) （20133011）
