Abstract: An improved normalized block semi-norm algorithm is presented. The mode of sampling contraction for the codebook and search for the nearest farther block are improved. It can reduce the searching range of the matched block and improve the matched accuracy, and in order to improve the coding rate and decoding image quality. Experimental results show that the coding time of the improved algorithm is shorter than the fast algorithm based on normalized block semi-norm.

Key words: fractal image coding, Human Visual System(HVS), quadtree partition, semi-norm

摘要： 提出一种改进的规范块半范数算法，对图像分割策略、码本块抽样收缩方式及搜索最近意义父块方式进行改进，以缩小子块搜索其最佳匹配父块的范围，加快编码速度，提高匹配准确度和解码图像质量。实验结果表明，与基于规范块半范数的快速算法相比，该算法的编码时间更短。

关键词: 分形图像编码, 人类视觉系统, 四叉树分割, 半范数

CLC Number: 

• TN919.81