摘要
为避开常规铅笔画生成算法中图像分割的限制,提出一种基于纹理与边缘轮廓的铅笔素描画自动生成算法.该算法先用主成分分析法对原始图像进行分析,确定高频提升过滤的参数,然后融合相应的相似度测量方法来检测并定位图像的边缘.同时,利用线积分卷积产生原始图像的纹理,并将其与边缘图叠加,得到铅笔素描画的效果.实验结果表明,该方法简单且有效.
To avoid the drawbacks of image segmentation used in many pencil drawing researches,we propose a new solution to generate pencil drawing based on texture and profile.Principal component analysis(PCA) is used to analyze the original image,and parameters of high-boost filter(HBT) determined.Edges in the image are located by combining some similarity measures.Line integral convolution(LIC) is used to generate the original image's texture.Pencil drawing is finally created by superposition of the edge image and the texture image.Experimental results show that the proposed method is simple and effective.
出处
《上海大学学报(自然科学版)》
CAS
CSCD
北大核心
2010年第3期312-317,共6页
Journal of Shanghai University:Natural Science Edition
基金
上海市重点学科建设资助项目(J50103)
上海市教委科研创新资助项目(09YZ10)
关键词
纹理
轮廓
铅笔素描画
texture
profile
pencil drawing