摘要
嵌入式零树小波(EZW)是一种非常有效的基于离散小波变换的图像编码算法,可以实现渐进编解码,具有较好的图像恢复质量。在研究嵌入式零树小波编码算法及原理的基础上,针对其对系数重复扫描而带来的计算量与编码比特数的增加,以及扫描过程中出现的大量零数根而导致的算法复杂度增加等不足,提出一种将改进的EZW算法与霍夫曼编码方法相结合的图像压缩编码方法。首先通过扩充编码符号改变扫描方式,来实现零树结构的快速判断,避免连续出现零数根。然后将改进的算法与霍夫曼编码联合编码来代替算术编码方法使其更简单。最后,在Matlab中模拟仿真,将改进算法与原算法进行比较分析。实验结果表明,与独立的EZW算法相比,改进算法不仅增加了编码效率,同时也提高了峰值信噪比(PSNR),证明了改进算法的有效可行性。
Embedded zerotree wavelet（ EZW） is a kind of very effective image coding algorithm based on discrete wavelet transform,which can achieve progressive code with a good quality of image recovery. In the research of embedded zero tree wavelet coding algorithm,there exist deficiency that repeated scanning of coefficient would increase the amount of calculation and the coding bits number,furthermore during the scanning process a large number of zero would result in high-order complexity.For this,we propose an improved EZW algorithm of image compression coding combing Huffman coding. First of all,it expands encoding dictionary to change scanning way,implementing the quick determination of the zero tree structures,avoiding continuous zero root.Then it is combined with the Huffman code to replace the arithmetic coding method to make it simpler.Finally,we compare the original algorithm with the proposed algorithm in the simulation with Matlab.The experiments showthat in comparison to the independent EZW algorithm,the proposed algorithm not only increases the coding efficiency,but also improves the peak signal-to-noise ratio（ PSNR）,which proves its feasibility.
作者
詹为
段先华
於跃成
ZHAN Wei;DUAN Xian-hua;YU Yue-cheng(School of Computer Science and Engineering,Jiangsu University of Science and Technology,Zhenjiang 212003,China)
出处
《计算机技术与发展》
2018年第6期21-25,共5页
Computer Technology and Development
基金
国家自然科学基金(61502211)
作者简介
詹为（1992-），女，硕士研究生，研究方向为图像处理；;段先华，教授，博士，研究方向为电子对抗、图像处理、模式识别等；;於跃成，副教授，博士，研究方向为机器学习、数据挖掘、模式识别等。
