摘要
针对传统基本分形编码存在的计算复杂性较高、编码时间较长的缺点,提出了一种基于双交叉和的特征值编码算法,以解决分形图像编码时间过长的问题。该算法通过构造图像块适当的特征向量,将"R在D集合中搜索MSE意义下的最佳匹配块"问题转换成"R的特征向量在D的特征向量空间中搜索最佳匹配块"的问题,将全局搜索转化为相对意义下的近邻搜索,使得匹配搜索只在初始匹配块的邻域内进行,有效地减少了搜索对象,从而进一步加快了编码速度。采用图像方块分割进行了多种算法的对比仿真实验,实验结果表明相对于其他算法,所提出的算法在保证一定重建图像质量的前提下,提高了图像的结构相似度,图像编码时间明显缩短,较好地实现了提高算法编码速度的目的。
For the shortcomings of high computational complexity and long encoding time of the traditional fractal coding, an encoding al- gorithm based on the sum of double cross eigenvalues is proposed in order to solve the problem of long encoding time. The problem, R search for the best matching block of MSE sense in the D set,is converted into another one, the eigenvector of R search for the best matc- hing block of D in the eigenvector space. Thus, global search is transformed into neighbor sea^h by constructing suitable feature vector for the image block in the opposite sense and matching search is carried out only in the field of the initial matching block, which reduces search objects and then speeds up encoding. Through comparative simulation experiment, a variety of algorithms are compared and simu-lated by using image segmentation. The results of experiments show that the algorithm presented has improved the feature similarity and reduces the image encoding time more effectively under the premise of ensuring quality of the reconstructed image, and that the purpose of improving the encoding speed procedure has been achieved.
出处
《计算机技术与发展》
2017年第3期159-162,共4页
Computer Technology and Development
基金
国家自然科学基金面上项目(11471114
61372125)
关键词
分形
分形图像编码
特征向量法
双交叉和特征
fractal
fractal image coding
eigenvector method
sum of double cross eigenvalues
