摘要: 针对图像非真实感绘制(NPR)算法和感兴趣区域(ROI)提取算法的不足,提出一种基于ROI的图像NPR算法。该算法利用Mean-Shift算法实现图像的NPR效果,结合图像的ROI解决传统算法时间效率不高的问题,应用色彩传递的方法对图像自动进行色彩添加和变更,以改善视觉效果。实验结果表明,该算法具有良好的艺术效果,能有效提高算法效率。
关键词:
感兴趣区域,
非真实感绘制,
Mean-Shift算法,
分水岭变换,
色彩传递
Abstract: Aiming at the shortcomings of image Non-Photorealistic Rendering(NPR) algorithms and Regions of Interest(ROI) extraction algorithms, this paper proposes a NPR algorithm based on ROI. It uses Mean-Shift algorithm to implement NPR effect, combines ROI of image to solve the problem of low-efficiency of existing algorithms, and uses the method of color transfer to reach a better visual effect. Experimental results show that the algorithm gains good art effect, and improves the efficiency.
Key words:
Regions of Interest(ROI),
Non-Photorealistic Rendering(NPR),
Mean-Shift algorithm,
watershed transform,
color transfer
中图分类号:
程琳琳, 陈昭炯. 基于感兴趣区域的图像非真实感绘制算法[J]. 计算机工程, 2010, 36(16): 195-197.
CHENG Lin-Lin, CHEN Zhao-Jiong. Image Non-Photorealistic Rendering Algorithm Based on Regions of Interest[J]. Computer Engineering, 2010, 36(16): 195-197.