Research On Color Stippling Based On Incremental Voronoi Sequence

Image stylization technology is a branch of computer graphics.In the pass twenty years,many researchers have explored Image stylization technology.It mainly used computer to give a certain style of image and make the generated images have artistic sense.Digital stippling as one of the most popular stylization techniques has been an attractive research topicRecently,most of the existing works focus on mono-tone stippling.The image of generation has many problems.such as,poor color of image,lack of visual rationality.Color stippling based on multi-class sampling is extremely computationally expensive.we present a novel algorithm for instant color stippling,including the following:(1)Our approach employs an incremental Voronoi set which uses a pre-computed sample sequence of blue noise properties.We extended the thresholding strategy for stipple generation to multi-classes cases to determine the color of sampling point and based on which can produce color stippling results.(2)Stippling with dots of different radius.Each pixel hue has been controlled the radius of dots and a threshold strategy is used to control the color of the dots.(3)The target detection and segmentation method based on neural network is used to draw the video.The target and segmentation are obtained by Mask R-CNN.Only place dots where the mask exists in each frame of the video.Compared with other composition strategies in different color space,the proposed approach can produce color stippling results with comparable quality to the state of the art while keeping the efficiency and robust.Color stippling results generated in video are also comparable to the quality of the color stippling generated by frame-by-frame,but the time efficiency is high.