首页> 外文会议>IEEE International Conference on Digital Signal Processing >Structurally orthogonal finite precision FPGA implementation of block-lifting-based quaternionic paraunitary filter banks for L2L image coding
【24h】

Structurally orthogonal finite precision FPGA implementation of block-lifting-based quaternionic paraunitary filter banks for L2L image coding

机译:用于L2L图像编码的基于块升降的季屈氏杠杆式滤波器的结构正交有限精度FPGA实现

获取原文

摘要

This paper presents a systematic design of the integer-to-integer invertible quaternionic multiplier based on the block-lifting structure and pipelined embedded processor of the given multiplier using distributed arithmetic (DA) as a block of M-band linear phase paraunitary filter banks (LP PUFB) based on the quaternionic algebra (Q-PUFB). A bank Q-PUFB based on the DA block-lifting structure reduces the number of rounding operations and has a regular layout, and can be implemented on the FPGA-based devices as integer-to-integer transform. Compared to general-purpose Q-PUFB obtained using lifting factorization, it needs less lifting steps and the given DA-based Q-PUFBs have less number of rounding operations and this property is useful for the lossless-to-lossy (L2L) image coding. Furthermore our DA-based 8-channel LP Q-PUFBs presented superior coding results on the PSNR than the corresponding filter banks, especially for image with relatively strong highpass components.
机译:本文介绍了基于给定乘法器的块提升结构和流水线嵌入式处理器的整数到整数可逆性倍增器的系统设计,该乘法器使用分布式算术(DA)作为M带线性相位占领滤波器组的块( LP PUFB)基于四元数代数(Q-PUFB)。基于DA块提升结构的BANK Q-PUFB减少了舍入操作的数量并具有常规布局,可以在基于FPGA的设备上实现为整数到整数变换。与使用升降分解获得的通用Q-PUFB相比,它需要较少的提升步骤,并且给定的基于DA的Q-PUFB具有较少数量的舍入操作,并且该属性对于无损损耗(L2L)图像编码是有用的。此外,基于DA的8通道LP Q-PUFBS在PSNR上呈现出比相应的滤波器组的卓越编码结果,特别是对于具有相对强大的高通组件的图像。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

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