摘要
基于image quilting纹理传输算法原理,研究纹理图像的梯度结构信息、源纹理图像与目标图像亮度误差对纹理传输效果的影响。在选择候选纹理块时,以纹理块的梯度结构信息与颜色误差同时作为判断标准。对源纹理图像进行亮度重映射,使其与目标图像的亮度处于相同尺度,可避免因亮度误差过大而导致无法获取最优块的情况。实验证明,与传统纹理传输算法相比,改进后的算法能取得更好的传输效果。
Based on image quilting texture transfer algorithm, we study the effects of texture gradient structure information and brightness error between source texture image and target image on texture transmission. When selecting the candidate texture block, we take both the gradient structure information and color error of the texture block as the criterion. Meanwhile, we remap the brightness of the source texture image to the same level as the brightness of the target image, which can avoid the case where the optimal block cannot be selected due to the too large brightness error. The experimental results show that compared with the traditional texture transfer algorithm, the improved algorithm can achieve a better transmission effect.
作者
谭永前
曾凡菊
Tan Yongqian1, Zeng Fanju1,2(1.School of Big Data Engineering, Kaili University, Kaili, Guizhou 556011, China; 2 College of Optoelectronic Engineering, Chongqing University, Chongqing 400044, Chin)
出处
《激光与光电子学进展》
CSCD
北大核心
2018年第3期180-184,共5页
Laser & Optoelectronics Progress
基金
贵州省教育厅青年科技人才成长项目[黔教合KY字(2017)335]
贵州省科技合作计划[黔科合LH字(2017)7161号
黔科合LH字(2016)7318号
黔科合LH字(2015)7744号]
国家自然科学基金(11464023)
