首页> 外国专利> Coding system and method for lossless and lossy compression of still and motion images

Coding system and method for lossless and lossy compression of still and motion images

机译:用于静止和运动图像的无损和有损压缩的编码系统和方法

摘要

The invention relates to a coding system for lossless and lossy compression of still and motion images with means of statistical context modeling for adaptive entropy coding of wavelet coefficients in different conditioning templates. Furthermore, it is directed to a method for lossless and lossy compression of still and motion image data by hierarchical decomposing of said image data into subbands by a revertible wavelet transform that generates wavelet coefficients and by using a conditioning template for statistical context modeling and adaptive entropy coding of said wavelet coefficients. The compression ratio is improved by one or more of the following steps:adapting the shape and/or orientation of said conditioning template to different subbands;converting a two-dimensional array of signed wavelet coefficients into an equivalent sequence of only two input symbols for adaptive binary entropy coding;using previously scanned bit planes in forming conditioning templates;reducing the number of possible conditioning states corresponding to all possible combinations of events in the conditioning templates by using least-squares estimates of magnitudes of wavelet coefficients;reducing the number of possible conditioning states corresponding to all possible combinations of events in the conditioning templates by first using least-squares estimates of magnitudes of wavelet conditions: and then by minimum-entropy quantization of said estimates:comparing the so-far-coded bits of the coefficient being presently coded with the so-far-coded bits of the neighbouring coefficients and parent coefficient of to characterise spacial texture patterns and using them to augment the conditioning states created by the quantization of said estimates;conditioning the sign of a wavelet coefficient on the signs of neighbouring coefficients of .recording for each subband the location of the most significant bit of the coefficient of maximum magnitude in the subband and including it as side information in the code stream.
机译:本发明涉及一种利用统计上下文建模的方法对静止和运动图像进行无损和有损压缩的编码系统,用于对不同条件模板中的小波系数进行自适应熵编码。此外,本发明涉及一种用于静止和运动图像数据的无损和有损压缩的方法,该方法是通过可逆小波变换将所述图像数据分层分解为子带,该可逆小波变换产生小波系数,并且通过使用条件模板进行统计上下文建模和自适应熵。所述小波系数的编码。通过以下一个或多个步骤可以提高压缩率: 使所述条件模板的形状和/或方向适应于不同的子带; 转换一个二维数组带符号的小波系数变成只有两个输入符号的等效序列,用于自适应二进制熵编码; 在形成条件模板时使用先前扫描的位平面; 减少对应的可能条件状态的数量通过使用小波系数幅度的最小二乘估计估计条件模板中事件的所有可能组合; 通过首先使用减少与条件模板中事件的所有可能组合相对应的可能条件状态的数量小波条件的大小的最小二乘估计:然后通过对所述估计的最小熵量化: 比较迄今编码的比特o f系数目前已使用so编码-Image图像的相邻系数和父系数的-far编码位=“ IMGA0001.GIF” he =“ 3” id =“ ia02” imgContent =“ character” imgFormat =“ GIF” inline =“ yes” wi =“ 2“ />来表征空间纹理图案,并使用它们来增强由所述估计的量化所创建的条件状态; <​​/ ListItem> 调节小波系数的符号关于<图像文件=“ IMGA0001.GIF”的相邻系数的符号he =“ 3 “ id =” ia04“ imgContent =” character“ imgFormat =” GIF“ inline =” yes“ wi =” 2“/>。 记录每个子带的系数最高有效位的位置在子带中具有最大幅度,并将其作为辅助信息包含在码流中。

著录项

  • 公开/公告号EP0914004A1

    专利类型

  • 公开/公告日1999-05-06

    原文格式PDF

  • 申请/专利权人 NTEC MEDIA GMBH;

    申请/专利号EP19970118846

  • 发明设计人 WU XIAOLIN;BARTHEL KAI UWE;

    申请日1997-10-29

  • 分类号H04N7/26;

  • 国家 EP

  • 入库时间 2022-08-22 02:19:01

相似文献

  • 专利
  • 外文文献
  • 中文文献
获取专利

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号