A Fast Computational Algorithm for the Discrete Cosine Transform

学习报告:A Fast Computational Algorithm for the Discrete Cosine Transform

基本信息

标题:A Fast Computational Algorithm for the Discrete Cosine Transform.
摘要:A Fast Discrete Cosine Transform algorithm has been developed which provides a factor of six improvement in computational complexity when compared to conventional Discrete Cosine Transform algorithms using the Fast Fourier Transform. The algorithm is derived in the form of matrices and illustrated by a signal-flow graph, which may be readily translated to hardware or software implementations.
关键词:Transforms; Computational complexity; Discrete Fourier transforms; Discrete cosine transforms;Fast Fourier transforms;Matrix decomposition.

中心思想

Abstract摘要、Introduction引言部分

该论文旨在介绍一种更快捷有效的离散余弦变换算法(DCT,Discrete Cosine Transform),与使用快速傅里叶变换(FFT,fast Fourier transform)的传统离散余弦变换算法相比,大大降低其计算复杂度,提高其运算速度。
离散余弦变换是最接近KL变换的正交变换,可有效地去除空间域的相关性并且可以使大部分的能量集中在直流和低频的部分,已成功应用于高分辨率图像的编码,其常规计算方法在整个计算过程中都采用了计算复杂度大的含有复数运算的双边快速傅里叶变换算法。由于缺乏有效的算法,离散余弦变换在各种应用中的用途并未得到充分开发。

Discrete Cosine Transform部分

该论文提出了一种快速离散余弦变换算法(FDCT,fast Discrete Cosine Transform),该算法以矩阵形式推导,并通过信号流图进行说明,以便后续轻松转换为硬件或软件实现。
该算法仅涉及N点的实数运算,并可以应用于满足 N = 2 m N=2^m N=2m(m不小于2)值的任意情况,主要思路概括如下。
将离散余弦变换表达式(1)表示成矩阵形式(2),通过矩阵理论中矩阵分解法 [ A N ] [A_N] [AN]分解,将离散函数 f ( j ) f(j) f(j)的离散余弦变换 F ( k ) F(k) F(k)由交替的余/正弦蝶形矩阵与二元矩阵表示,将矩阵元素重新排序,保留信号流图节点的位反转模式(bit-reversed)。
在这里插入图片描述

实现效果

(1)N点新FDCT算法需要 1.5 N l o g 2 N − 1 + 2 1.5Nlog_2N-1+2 1.5Nlog2N1+2次实数加法运算及 N l o g 2 N − 1.5 N + 4 Nlog_2N-1.5N+4 Nlog2N1.5N+4次实数乘法,是传统FDCT算法运算次数的1/6,是FFT算法运算次数的1/3;
(2)信号流图中输入样本值按自然顺序排列,输出变换系数按位反转(bit-reversed)顺序排列;
(3)将信号流图扩展到2的下一个幂的N值时,仅涉及添加一组正负1蝶形以容纳新的输入样本集;通过将先前FDCT变换系数的下标加倍以产生新的偶数变换系数集;添加一系列交替的余弦/正弦蝶形和正负1蝶形以产生新的奇数变换系数集。
(4)信号流图是双向的,对应着新FDCT变换和IFDCT变换。

句型积累

(1)The algorithm is derived in the form of …
(2)Use of the XXX in a wide variety of applications has not been as extensive as its properties would imply due to the lack of an efficient algorithm.
(3)The generalization is not unique-several alternate methods have been discovered but the method described herein appears to be the simplest to interpret.
(4)It is not necessarily the most efficient XXX which could be constructed but represents one technique for methodical extension.
(5)XXX can be expressed in a XXX form as …
(6)Note that…It can be seen that…

猜你喜欢

转载自blog.csdn.net/Echoshit8/article/details/113094322