您的位置:首页  > 论文页面

基于人眼视觉特性的图像自适应量化

发表时间:2012-11-30  浏览量:1057  下载量:362
全部作者: 马建林,马义德
作者单位: 兰州大学信息科学与工程学院;空军95061部队
摘 要: 利用人眼视觉特性的原理,提出一种基于脉冲耦合神经网络(pulse coupled neural network, PCNN)的图像量化算法,这是一种符合人眼视觉特性的非均匀量化方法,对比常见的均匀量化其能很好地提高图像量化算法的有效性。首先将待量化灰度图像导入实验选择的PCNN;然后根据输入图像的不同自适应地进行PCNN的参数配置,在参数选择完成后利用PCNN的性质进行图像的分层量化;最后对不同量化方式下得到的量化图像分别进行霍夫曼编码,通过对比不同量化方式下霍夫曼编码的码长,很好地证明了PCNN量化的有效性。
关 键 词: 信号与信息处理;图像量化;人眼视觉特性;脉冲耦合神经网络;霍夫曼编码
Title: Adaptive image quantization based on human visual characteristics
Author: MA Jianlin, MA Yide
Organization: School of Information Science and Engineering, Lanzhou University; Troop 95061 of Air Force
Abstract: Combined with the human visual characteristics, this paper describes a method for adaptive image quantization using pulse coupled neural network (PCNN). In the proposed method, a non-uniform quantization approach is used to limit the steps of image quantization rather than uniform quantization. This increases the efficiency of the quantization algorithm. Furthermore, we extract information from an input gray image and attempt to build a direct relation between the dynamic properties of PCNN and the static properties of each input image, and then an automatic parameter setting method for PCNN is applied to make the image quantization algorithm even more intelligent. Finally, the different quantitative images are coded using Huffman coding, and the experimental results show that PCNN can be used to quantize images effectively.
Key words: signal and information processing; image quantization; human visual characteristics; pulse coupled neural network; Huffman coding
发表期数: 2012年11月第22期
引用格式: 马建林,马义德. 基于人眼视觉特性的图像自适应量化[J]. 中国科技论文在线精品论文,2012,5(22):2176-2182.
 
0 评论数 0
暂无评论
友情链接