Image Smoothing via L0 Gradient Minimization

晃晃

Image Smoothing via L0 Gradient Minimization

Li Xu∗ Cewu Lu∗ Yi Xu Jiaya Jia

Department of Computer Science and Engineering

The Chinese University of Hong Kong







Abstract

We present a new image editing method, particularly effective for

sharpening major edges by increasing the steepness of transition

while eliminating a manageable degree of low-amplitude s***ctures.

The seemingly contradictive effect is achieved in an optimization

framework making use of L0 gradient minimization, which can

globally control how many non-zero gradients are resulted in to

approximate prominent s***cture in a sparsity-control manner. Un-

like other edge-preserving smoothing approaches, our method does

not depend on local features, but instead globally locates impor-

tant edges. It, as a fundamental tool, finds many applications and

is particularly beneficial to edge extraction, clip-art JPEG artifact

removal, and non-photorealistic effect generation.

Keywords: image smoothing, L0 sparsity, sharpening, filtering

1 Introduction

Photos comprise rich and well-s***ctured visual information. In

human visual perception, edges are effective and expressive stimu-

lation, vital for neural interpretation to make the best sense of the

scene. In manipulating and understanding pictures, high-level in-

ference with regard to salient s***ctures was intensively attended

to. Research following this line embodies generality and useful-

ness in a wide range of applications, including image recognition,

segmentation, object classification, and many other photo editing

and non-photorealistic rendering tasks.







We in this paper present a new editing tool, greatly helpful

for characterizing and enhancing fundamental image constituents,

i.e., salient edges, and in the meantime for diminishing insignif-

icant details. Our method relates in spirit to edge-preserving

smoothing [Tomasi and Manduchi 1998; Durand and Dorsey 2002;

Paris and Durand 2006; Farbman et al. 2008; Subr et al. 2009;

Kass and Solomon 2010] that aims to retain primary color change,

and yet differs from them in essence in focus and in mechanism.

Our objective is to globally maintain and possibly enhance the most

prominent set of edges by increasing steepness of transition while

not affecting the overall acutance. It enables faithful principal-

s***cture representation.

Algorithmically, we propose a sparse gradient counting scheme in

an optimization framework. Themain contribution is a new strategy

to confine the discrete number of intensity changes among neigh-

boring pixels, which links mathematically to the L0 norm for in-

formation sparsity pursuit. This idea also leads to an unconven-

tional global optimization procedure involving a discrete metric,

whose solution enables diversified edge manipulation according to

saliency. The qualitative effect of our method is to thin salient

edges, which makes them easier to be detected and more visually

distinct. Different from color quantization and segmentation, our

enhanced edges are generally in line with the original ones. Even

small-resolution objects and thin edges can be faithfully maintained

if they are s***cturally conspicuous, as shown in Fig. 1.









全文请下载附件：

lanshe

I'm eager for the code of realizing this formulation.Please send to me(631227964@qq.com).   

Thank you very much!

lanshe

跪求 急求高手赐予代码············

C.R.CAN

楼主收集的可真全哦

tc

“再次路过……”我造一个-----特别路过

tc

佩服，好多阿 ，哈哈

菜刀吻电线

不错哦,顶一下......

菜刀吻电线

不错哦,顶一下......

晃晃

凡系斑竹滴话要听；凡系朋友滴帖要顶

奇

加精、加亮滴铁子，尤其要多丁页丁页