High-Simulation Oil-Painting-Style Images Based on Stroke Characteristics

Abstract

Stylised rendering is a research field that combines computer technology and painting art. It simulates an artist’s effect via a computer simulation to generate different artistic styles rapidly. Two methods based on graphic image processing can be used for the computer simulation of oil painting style, and these two are simulation of art media, such as oil painting canvas and brush, and simulation of the art creation process to produce oil paintings automatically. This study aims to combine these two methods and fully consider the image tone reproduction of original oil paintings and the representative colours of each painter’s oil painting to simulate the effect of the artist’s creation. The brush texture of a painter is applied to generate stylised images through a non-realistic rendering algorithm. The tone reproduction (greyscale histogram) of digital photos is adjusted to fit the tone reproduction of the target oil painting (grayscale histogram) and make the produced oil painting style image approximate the artist’s creation. Then, the hue of the digital photos is adjusted, and the colour and appearance of the painter’s oil painting are imitated from the hue. Through these steps, the final rendering effect of the stylisation treatment of oil paintings and the similarity with the target image are improved. The overall effect on landscape painting is better than that on portrait painting.