Oil Painting Style Rendering Based on Kuwahara Filter

Junjie Gao,Dajin Li,Wenran Gao

doi:10.1109/access.2019.2931037

Abstract

The Kuwahara filter can generate local smoothed color regions similar to hand-painted strokes, and therefore it can be used to image artistic rendering. However, the existing Kuwahara filtering methods often adopt a fixed-sized filter kernel, which leads to a homogeneous size of strokes. In view of this problem, a multi-scale anisotropic Kuwahara filtering method is proposed in this paper. By using image saliency as control information to adjust the size of filter kernel, strokes with abundant varied geometric scales are generated. This paper also presents an automatic rendering framework based on the proposed algorithm that automatically transforms images into an oil painting style. Using the image's orientation field of edge, edge gradient magnitude and image saliency to control the bump mapping, the senses of thickness and layers of oil painting's strokes are generated. By adopting this framework, an artistic oil painting style similar to the “impasto technique” can be simulated well.

Highlights

INTRODUCTIONPainters discard some of the inessential details, adopt fewer strokes to show the object’s main structure features (e.g., feature lines and object contours), and use simplified color details to fill the object’s internal region
In some painting styles, painters discard some of the inessential details, adopt fewer strokes to show the object’s main structure features, and use simplified color details to fill the object’s internal region
Image stylization methods can be classified into three categories: image artistic abstraction; stroke-based rendering; and neural style transfer (NST) which was proposed in recent years

Summary

INTRODUCTION

Painters discard some of the inessential details, adopt fewer strokes to show the object’s main structure features (e.g., feature lines and object contours), and use simplified color details to fill the object’s internal region. Based on the Kuwahara filter of round kernel, Kyprianidis et al [6] proposed a kind of anisotropic Kuwahara filter, forming short bar-shaped color blocks along image edge tangent. All the previous oil painting texture generation methods based on bump mapping did not consider the pigment thickening effect of stroke contours, which causes a lack of the sense of layers among the strokes. A filter-based image stylization method for oil painting style is proposed This method can generate the senses of thickness and the layers of oil painting strokes. Combining the edge tangent flow map, edge gradient magnitude, and saliency of the image to control the bump mapping to generate the stroke texture of an oil painting It can generate the effect of thickening pigment at the edge of strokes, generating a stronger sense of layers. It can generate strokes with different senses of thickness in different image regions

RELATED WORKS

STROKE GENERATION BASED ON KUWAHARA FILTERING

MULTI-SCALE ANISOTROPY KUWAHARA FILTERING

RESULTS AND DISCUSSION

CONCLUSION