Scale and Gradient Aware Image Smoothing

Shuai Fang,Jing Zhang,Zhenji Yao

doi:10.1109/access.2019.2953550

Shuai Fang, Jing Zhang + Show 1 more

Open Access

https://doi.org/10.1109/access.2019.2953550

Copy DOI

Abstract

This paper presents a novel scale and gradient aware image smoothing method, particularly effective for removing high-contrast detailed textures while preserving boundary sharpness and fine structures. The core idea of the proposed method is based on an observation that small-scale textures can be removed only depending on a down-then-up scaling(DTUS) operation. Accordingly, we present a selective edge smoothing framework by jointly considering scale and gradient measurements. Specifically, we first compute the structural similarity(SSIM) of the DTUS image and the input image to distinguish the textures and structures from the input image. Then we use the SSIM map as weights to fuse the two images together to achieve a scale aware smoothing result. Furthermore, we use the fusion image as guidance to confine the number of non-zero gradients and perform a guided L0 gradient minimization to achieve gradient aware smoothing. Since our method makes full use of image scale and gradient, it outperforms state-of-the-art image smoothing algorithm, especially in removing high-contrast textures. Since our proposed method can remove the insignificant details while preserving sharp and undistorted structural edges, it is also adaptable to many application scenarios, such as boundary extraction, image abstraction, JPEG artifact removal, and layer decomposition-based editing.

Highlights

Image smoothing is an important technique that removes detailed textures or noises while maintaining prominent edges
This paper presents a novel scale and gradient aware image smoothing method based on a scale-guided L0 gradient minimization framework
We present a guided L0 gradient minimization framework to preserve the steepness of prominent structural edges without affecting the global acutance

Summary

INTRODUCTION

Image smoothing is an important technique that removes detailed textures or noises while maintaining prominent edges. Since our method makes full use of image scale and gradient, it can effectively remove highcontrast textures while preserving structural edges. (1) A simple but effective scale aware strategy, down-thenup scaling (DTUS), is proposed to distinguish the textures and structures from a single image. This strategy can effectively detect high-contrast textures regardless of their gradient magnitudes. The image scaling strategy works well in the overlapping regions of structures and textures, as shown in the red rectangles of Fig.2 This observation illustrates that the DTUS operation can effectively remove the detailed texture regardless of its contrast or strength.

SSIM MAP GENERATION

SCALE-GUIDED L0 GRADIENT MINIMIZATION

SOLVING OBJECTIVE FUNCTION

APPLICATIONS

Findings

CONCLUSION