摘要
本文提出了一种新颖的滤波方法—局部线性全变差滤波方法(LLTV),可以对数字图像进行有效的滤波处理.该方法采用图像的局部线性近似模型,并结合了全变差规则项,在滤除噪声的同时,能够很好的保持图像的边缘信息,而且其简单的、线性时间算法,可以对各种数据进行快速处理.实验结果表明,在计算机视觉的各种应用中,如去噪、平滑、细节增强和高动态范围(HDR)压缩等方面,该方法都表现出高效的性能.
This paper proposes a novel edge-preserving smoothing filter-local linear total variation fdter （LLTV）, used to perform effective image filtering. Based on a local linear model and combining total variation regularization, this method can filter out noise while preserving edges and fine-scale details. The experimental results show that LLTV filter is effective in a great variety of computer vision applications including noise reduction,detail smoothing/enhancement and HDR compression.
出处
《电子学报》
EI
CAS
CSCD
北大核心
2012年第12期2507-2511,共5页
Acta Electronica Sinica
基金
国家自然科学基金(No.61139001
No.61172108
No.81241059)
关键词
局部线性模型
全变差
细节增强
HDR压缩
local linear model
total variation
detail enhancement
high dynamic range compression
