吉林大学学报(理学版)

• 计算机科学 • 上一篇    下一篇

计算全息图的小波神经网络压缩方法

侯阿临, 吴亮, 廖庆, 王崇锦, 郭俊良   

  1. 长春工业大学 计算机科学与工程学院, 长春 130012
  • 收稿日期:2014-08-04 出版日期:2015-07-26 发布日期:2015-07-27
  • 通讯作者: 吴亮 E-mail:wuliangdesky@126.com

Hologram Compression by Wavelet Neural Network

HOU Alin, WU Liang, LIAO Qing, WANG Chongjin, GUO Junliang   

  1. College of Computer Science and Engineering, Changchun University of Technology, Changchun 130012, China
  • Received:2014-08-04 Online:2015-07-26 Published:2015-07-27
  • Contact: WU Liang E-mail:wuliangdesky@126.com

摘要:

采用小波神经网络方法对信息量较大、 难提高压缩效率的计算全息图进行数据压缩, 利用其较强的非线性映射和函数逼近能力自适应地调整和处理全息图, 可大幅减少信息冗余, 得到较好的压缩效果. 实验结果表明: 利用该算法能得到1.56%的低压缩率, 此时的再现像较清晰, 失真较小; 与常用压缩算法相比, 当压缩率很低时, 用小波神经网络压缩全息图是一种切实可行且更有效的方法.

关键词: 计算全息, 图像压缩, 小波分析, 小波神经网络

Abstract:

The algorithm of wavelet neural network was presented to compress the data of computergenerated hologram  which has much more information and is difficult to raise compression rate. The method can adaptively adjust and process the nonlinear hologram with a strong ability of nonlinear mapping and functional approximation, so that it can greatly reduce the redundancy of information. The obtained compression rate is as low as 156% in the experiments, and at this time the reconstructed image is relatively clear with small distortion. Compared with several current compression algorithms, the proposed hologram compression method of wavelet neural network is feasible and more effective with regard to those holograms with low compression rate.

Key words: computergenerated hologram, image compression, wavelet analysis, wavelet neural network

中图分类号: 

  • TP18