Computer Engineering and Applications ›› 2007, Vol. 43 ›› Issue (11): 45-47.

• 学术探讨 • Previous Articles     Next Articles

ROI Coding Based On Image Characteristic Quantisation

  

  • Received:2006-05-15 Revised:1900-01-01 Online:2007-04-11 Published:2007-04-11

基于图像特征量化的感兴趣区域编码

李景超 罗建书   

  1. 国防科大理学院 国防科学技术大学数学理学院五队
  • 通讯作者: 李景超

Abstract: Using the fact that the wavelet transform matches the HVS, this article presents a commingled quantization encoding method based on various wavelet straps’characters and pertinence within the same wavelet strap.According to the encoding method, it can support progressive transmission and with the experiment, it provides satisfied visual reconstructed image in high compression ratio.

Key words: image characteristic quantisation, HVS, EZW algorithm, vector quantization, ROI coding

摘要: 本文利用小波变换中多分辨率分析与人眼视觉匹配的特性,针对图像在小波域不同子带中的特点及重要性,提出了一种基于子带特征,结合带间与带内相关性进行混合量化编码的方法,支持渐进式传输,并在高压缩比下获得了视觉上较满意的高质量恢复图像。

关键词: 图像特征量化, 人眼视觉系统, EZW算法, 矢量量化, 感兴趣区域编码