Golden section transform
- This article was considered for deletion at Wikipedia on December 22 2013. This is a backup of Wikipedia:Golden section transform. All of its AfDs can be found with PrefixIndex, the first at Wikipedia:Wikipedia:Articles for deletion/Golden section transform. If the page name here has changed, please see these three pages instead.
- Wikipedia editors had multiple issues with this page:
- The topic of this article may not meet Wikipedia's general notability guideline. But, that doesn't mean someone has to… establish notability by citing reliable secondary sources that are independent of the topic and provide significant coverage of it beyond its mere trivial mention. (November 2013)
Template:Confusing Template:Lead rewrite Template:Technical oooh, orphan
The golden section transform was proposed and used in digital image multiresolution analysis in 2007 by Jun Li.[1] The ratio of 1 level "Low" part length to "High" part length is the approximation of the golden ratio <math>\varphi</math> (Type L) or <math>\frac{1}{\varphi}</math> (Type H),
where
- <math>\varphi = \frac{1+\sqrt{5}}{2} = 1.618\ldots ;</math>
- <math>\frac{1}{\varphi}= \frac{-1+\sqrt{5}}{2} = 0.618\ldots</math>
Contents
DCT-golden section transform (DCT-GST)
It is well known that the Fibonacci sequence is:
- <math>0,\;1,\;1,\;2,\;3,\;5,\;8,\;13,\;21,\;34,\;55,\;89,\;144,\;233,\;377,\;610,\;987\ldots\;</math> Template:OEIS
- <math>F_0 = 0,\; F_1= 1,F_n = F_{n-1} + F_{n-2},</math> make a golden section tree of the Fibonacci number 8:
Fibonacci number | DCT-LGST | Reverse order | Symmetrical form | DCT-HGST | Reverse order | symmetrical form |
---|---|---|---|---|---|---|
1 | 1 | 1 | 1 | – | – | – |
2 | 2 | 2 | 2 | 2 | 2 | 2 |
3 | 12 | 21 | 21 | 3 | 3 | 3 |
5 | 212 | 212 | 212 | 23 | 32 | 32 |
8 | 12212 | 21221 | 21212 | 323 | 323 | 323 |
13 | 21212212 | 21221212 | 21212212 | 23323 | 32332 | 32323 |
21 | 1221221212212 | 2122121221221 | 2121221221212 | 32323323 | 32332323 | 32323323 |
DCT-low golden section transform (DCT-LGST)
Take a sequence <math>(p_0,p_1,p_2,p_3,p_4,p_5,p_6,p_7)</math> of 8 elements for example, we get the form of 8 is <math>12212</math>, 1-level DCT-LGST of the 8 elements is:
- <math>p_0,\;\frac{p_1 + p_2}{2},\;\frac{p_3 + p_4}{2},\;p_5,\;\frac{p_6 + p_7}{2},\;\frac{p_1 - p_2}{2},\;\frac{p_3 - p_4}{2},\;\frac{p_6 - p_7}{2}</math>
the "low" part of the result is:
- <math>p_0,\;\frac{p_1 + p_2}{2},\;\frac{p_3 + p_4}{2},\;p_5,\;\frac{p_6 + p_7}{2}</math>
according to the form <math>212</math> of 5, 2-level DCT-LGST of the 8 elements is:
- <math>\frac{p_0 + \frac{p_1 + p_2}{2}}{2},\;\frac{p_3 + p_4}{2},\;\frac{p_5 + \frac{p_6 + p_7}{2}}{2},\;\frac{p_0 - \frac{p_1 + p_2}{2}}{2},\;\frac{p_5 - \frac{p_6 + p_7}{2}}{2},\;\frac{p_1 - p_2}{2},\;\frac{p_3 - p_4}{2},\;\frac{p_6 - p_7}{2}</math>
Let <math>L_{N}</math> be the <math>N\times N</math> non-normalized DCT-LGST matrix, where <math>N=F_k, (k\ge3,k \in \mathbb{Z})</math>
when <math>N=2</math>, 1-level non-normalized DCT-LGST matrix is:
- <math>L_{2}=\frac{1}{2}\begin{bmatrix}
1 & 1\\1 & -1\\ \end{bmatrix}</math>
when <math>N=3</math>, 2-level non-normalized DCT-LGST matrix and the inverse matrix is:
- <math>L_{3}=\frac{1}{4}\begin{bmatrix}
2 & 1 & 1\\2 & -1 & -1\\0 & 2 & -2\\ \end{bmatrix}</math>; <math>L_{3}^{-1}=\begin{bmatrix} 1 & 1 & 0\\1 & -1 & 1\\1 & -1 & -1\\ \end{bmatrix}</math>
when <math>N=5</math>, 3-level non-normalized DCT-LGST matrix and the inverse matrix is:
- <math>L_{5}=\frac{1}{8}\begin{bmatrix}
2 & 2 & 2 & 1 & 1\\2 & 2 & -2 & -1 & -1 \\0 & 0 & 4 & -2 & -2\\4 & -4 & 0 & 0 & 0 \\0 & 0 & 0 & 4 & -4\\ \end{bmatrix}</math>; <math>L_{5}^{-1}=\begin{bmatrix} 1 & 1 & 0 & 1 & 0\\1 & 1 & 0 & -1 & 0 \\1 & -1 & 1 & 0 & 0\\1 & -1 & -1 & 0 & 1 \\1 & -1 & -1 & 0 & -1\\ \end{bmatrix}</math>
when <math>N=8</math>, 4-level non-normalized DCT-LGST matrix and the inverse matrix is:
- <math>L_{8}=\frac{1}{16}\begin{bmatrix}
4 & 2 & 2 & 2 & 2 & 2 & 1 & 1\\ 4 & 2 & 2 & -2 & -2 & -2 & -1 & -1\\ 0 & 0 & 0 & 4 & 4 & -4 & -2 & -2\\ 8 & -4 & -4 & 0 & 0 & 0 & 0 & 0\\ 0 & 0 & 0 & 0 & 0 & 8 & -4 & -4\\ 0 & 8 & -8 & 0 & 0 & 0 & 0 & 0\\ 0 & 0 & 0 & 8 & -8 & 0 & 0 & 0\\ 0 & 0 & 0 & 0 & 0 & 0 & 8 & -8 \end{bmatrix}</math>;<math>L_{8}^{-1}=\begin{bmatrix} 1 & 1 & 0 & 1 & 0 & 0 & 0 & 0\\ 1 & 1 & 0 & -1 & 0 & 1 & 0 & 0\\ 1 & 1 & 0 & -1 & 0 & -1 & 0 & 0\\ 1 & -1 & 1 & 0 & 0 & 0 & 1 & 0\\ 1 & -1 & 1 & 0 & 0 & 0 & -1 & 0\\ 1 & -1 & -1 & 0 & 1 & 0 & 0 & 0\\ 1 & -1 & -1 & 0 & -1 & 0 & 0 & 1\\ 1 & -1 & -1 & 0 & -1 & 0 & 0 & -1 \end{bmatrix}</math>
when <math>N=8</math>, 4-level non-normalized reverse-order DCT-LGST matrix and the inverse matrix is:
- <math>L_8=\frac{1}{16}\begin{bmatrix}
1 & 1 & 2 & 2 & 2 & 2 & 2 & 4\\ 1 & 1 & 2 & 2 & 2 & -2 & -2 & -4\\ 2 & 2 & 4 & -4 & -4 & 0 & 0 & 0\\ 4 & 4 & -8 & 0 & 0 & 0 & 0 & 0\\ 0 & 0 & 0 & 0 & 0 & 4 & 4 & -8\\ 8 & -8 & 0 & 0 & 0 & 0 & 0 & 0\\ 0 & 0 & 0 & 8 & -8 & 0 & 0 & 0\\ 0 & 0 & 0 & 0 & 0 & 8 & -8 & 0 \end{bmatrix}</math>;<math>L_{8}^{-1}=\begin{bmatrix} 1 & 1 & 1 & 1 & 0 & 1 & 0 & 0\\ 1 & 1 & 1 & 1 & 0 & -1 & 0 & 0\\ 1 & 1 & 1 & -1 & 0 & 0 & 0 & 0\\ 1 & 1 & -1 & 0 & 0 & 0 & 1 & 0\\ 1 & 1 & -1 & 0 & 0 & 0 & -1 & 0\\ 1 & -1 & 0 & 0 & 1 & 0 & 0 & 1\\ 1 & -1 & 0 & 0 & 1 & 0 & 0 & -1\\ 1 & -1 & 0 & 0 & -1 & 0 & 0 & 0 \end{bmatrix}</math>
also, when <math>N=8</math>, 4-level non-normalized symmetrical form DCT-LGST matrix and the inverse matrix is:
- <math>L_{8}=\frac{1}{16}\begin{bmatrix}
1 & 1 & 2 & 2 & 2 & 4 & 2 & 2\\ 1 & 1 & 2 & 2 & 2 & -4 & -2 & -2\\ 2 & 2 & 4 & -4 & -4 & 0 & 0 & 0\\ 4 & 4 & -8 & 0 & 0 & 0 & 0 & 0\\ 0 & 0 & 0 & 0 & 0 & 8 & -4 & -4\\ 8 & -8 & 0 & 0 & 0 & 0 & 0 & 0\\ 0 & 0 & 0 & 8 & -8 & 0 & 0 & 0\\ 0 & 0 & 0 & 0 & 0 & 0 & 8 & -8 \end{bmatrix}</math>;<math>L_{8}^{-1}=\begin{bmatrix} 1 & 1 & 1 & 1 & 0 & 1 & 0 & 0\\ 1 & 1 & 1 & 1 & 0 & -1 & 0 & 0\\ 1 & 1 & 1 & -1 & 0 & 0 & 0 & 0\\ 1 & 1 & -1 & 0 & 0 & 0 & 1 & 0\\ 1 & 1 & -1 & 0 & 0 & 0 & -1 & 0\\ 1 & -1 & 0 & 0 & 1 & 0 & 0 & 0\\ 1 & -1 & 0 & 0 & -1 & 0 & 0 & 1\\ 1 & -1 & 0 & 0 & -1 & 0 & 0 & -1 \end{bmatrix}</math>
DCT-high golden section transform (DCT-HGST)
Similar to DCT-LGST, 1-level DCT-HGST of the 8 elements by the form <math>323</math> is:
- <math>\frac{p_0 + p_1+p_2}{3},\;\frac{p_3 + p_4}{2},\;\frac{p_5+p_6 + p_7}{3},\;
\frac{\sqrt{6}}{6}p_0 - \frac{\sqrt{6}}{6}p_2,\;\frac{\sqrt{2}}{6}p_0 -\frac{\sqrt{2}}{3}p_1 +\frac{\sqrt{2}}{6}p_2,\;\frac{p_3 - p_4}{2} ,\;</math>
- <math>\frac{\sqrt{6}}{6}p_5 - \frac{\sqrt{6}}{6}p_7
,\;\frac{\sqrt{2}}{6}p_5 -\frac{\sqrt{2}}{3}p_6 +\frac{\sqrt{2}}{6}p_7</math>
DCT-HGST can be also defined by a <math>N\times N</math> matrix <math>H_{N}</math>, where <math>N=F_k, (k\ge3,k \in \mathbb{Z})</math>, it is known that the <math>3\times 3</math> non-normalized DCT matrix is:
- <math>W=\frac{1}{3}
\begin{bmatrix} 1 & 1& 1 \\ \frac{\sqrt{6}}{2} &0&-\frac{\sqrt{6}}{2} \\ \frac{\sqrt{2}}{2} &-\sqrt{2}& \frac{\sqrt{2}}{2} \end{bmatrix} </math>
when <math>N=8</math>, 2-level non-normalized DCT-HGST matrix and the inverse matrix is:
- <math>H_{8}=\frac{1}{18}\begin{bmatrix}
2 & 2 & 2 & 3 & 3 & 2 & 2 & 2\\ \sqrt{6}&\sqrt{6}&\sqrt{6}& 0 & 0 & -\sqrt{6}&-\sqrt{6}&-\sqrt{6}\\ \sqrt{2}&\sqrt{2}&\sqrt{2}& -3\sqrt{2}& -3\sqrt{2}& \sqrt{2}&\sqrt{2}&\sqrt{2}\\ 3\sqrt{6} & 0 & -3\sqrt{6} & 0 & 0 & 0 & 0 & 0\\ 3\sqrt{2}&-6\sqrt{2}&3\sqrt{2}& 0 & 0 & 0 & 0 & 0\\ 0 & 0 & 0 & 9 & -9 & 0 & 0 & 0\\ 0 & 0 & 0 & 0 & 0 & 3\sqrt{6} & 0 & -3\sqrt{6}\\ 0 & 0 & 0 & 0 & 0 & 3\sqrt{2}&-6\sqrt{2}&3\sqrt{2}\\ \end{bmatrix};</math>
- <math>H_{8}^{-1}=\begin{bmatrix}
1 & \frac{\sqrt{6}}{2} & \frac{\sqrt{2}}{2} & \frac{\sqrt{6}}{2} & \frac{\sqrt{2}}{2} & 0 & 0 & 0\\ 1 & \frac{\sqrt{6}}{2} & \frac{\sqrt{2}}{2} & 0 & -\sqrt{2} & 0 & 0 & 0\\ 1 & \frac{\sqrt{6}}{2} & \frac{\sqrt{2}}{2} & -\frac{\sqrt{6}}{2} & \frac{\sqrt{2}}{2} & 0 & 0 & 0\\ 1 & 0 & -\sqrt{2} & 0 & 0 & 1 & 0 & 0\\ 1 & 0 & -\sqrt{2} & 0 & 0 & -1 & 0 & 0\\ 1 & -\frac{\sqrt{6}}{2} & \frac{\sqrt{2}}{2} & 0 & 0 & 0 & \frac{\sqrt{6}}{2} & \frac{\sqrt{2}}{2}\\ 1 & -\frac{\sqrt{6}}{2} & \frac{\sqrt{2}}{2} & 0 & 0 & 0 & 0 & -\sqrt{2}\\ 1 & -\frac{\sqrt{6}}{2} & \frac{\sqrt{2}}{2} & 0 & 0 & 0 & -\frac{\sqrt{6}}{2} & \frac{\sqrt{2}}{2} \end{bmatrix}</math>
DFT-golden section transform (DFT-GST)
Similarly, DFT-golden section transform can be defined as well, the <math>2\times 2</math> DFT matrix is identical with <math>2\times 2</math> DCT matrix, the <math>3\times 3</math> non-normalized DFT matrix is:
- <math>
W=\frac{1}{3} \begin{bmatrix} 1 & 1& 1 \\ 1 &-\frac{1}{2} - \frac{\sqrt{3}}{2}i&-\frac{1}{2} + \frac{\sqrt{3}}{2}i \\ 1 &-\frac{1}{2} + \frac{\sqrt{3}}{2}i& -\frac{1}{2} - \frac{\sqrt{3}}{2}i \end{bmatrix} </math>
The lifting scheme of reverse-order DCT-LGST
Let <math>F_0 = 0,\; F_1= 1,</math> let <math>\{X_u\}\,\!</math> be the original sequence, let <math>\{S_v\}\,\!</math> be a sequence as the "Low" part of reverse-order DCT-LGST, let <math>\{D_w\}\,\!</math> be a sequence as the "High" part of reverse-order DCT-LGST, we have
- <math>\{X_u\}=X_0, X_1, X_2, \ldots, X_u;\;</math> <math>\{S_v\}=S_0, S_1, S_2, \ldots, S_v;\;</math> <math>\{D_w\}=D_0, D_1, D_2, \ldots, D_w\;</math>
where <math>u=F_k-1,\;v=F_{k-1}-1,\;w=F_{k-2}-1\;(k\ge3,k \in \mathbb{Z})</math>
The lifting scheme of normalized reverse-order DCT-LGST is:
- <math>\begin{cases}S_{M_n}^{0} = X_{L_n-1},\; D_{n-1}^{0} = X_{L_n}\\
D_{n-1}^{1} = S_{M_n}^{0}-D_{n-1}^{0},S_{M_n}^{1} = S_{M_n}^{0}-0.5D_{n-1}^{1}\\S_{M_n} = \sqrt{2}S_{M_n}^{1}, \;S_{L_m} = X_{H_m}, \;D_{n-1}=D_{n-1}^{1}/\sqrt{2}\end{cases}</math>
and the reconstruction algorithm is:
- <math>\begin{cases}S_{M_n}^1 = S_{M_n}/\sqrt{2},\;D_{n-1}^1 =\sqrt{2}D_{n-1}
\\S_{M_n}^0 = S_{M_n}^1+0.5D_{n-1}^1,\;D_{n-1}^0 = S_{M_n}^0-D_{n-1}^1 \\X_{L_n-1}= S_{M_n}^0,\;X_{H_m}=S_{L_m},\;X_{L_n}=D_{n-1}^0 \end{cases}</math>
where
- <math>\{L_n\}=1, 4, 6, 9, 12, 14, 17, 19, \ldots\;</math> Template:OEIS
- <math>\{M_n\}=0, 2, 3, 5, 7, 8, 10, 11, \ldots\;</math> Template:OEIS
- <math>\{H_n\}=2, 7, 10, 15, 20, 23, 28, 31, \ldots\;</math> Template:OEIS
- <math>L_n = \left\lfloor \frac{3+\sqrt{5}}{2}n\right\rfloor-1
, M_n = \left\lfloor \frac{1+\sqrt{5}}{2}n\right\rfloor-1 , H_n = 2\left\lfloor \frac{1+\sqrt{5}}{2}n\right\rfloor+n-1</math>
where <math>m=1, 2, 3, 4, \ldots\;F_{k-3};\;n=1, 2, 3, 4, \ldots\;F_{k-2}\;(m,n \in \mathbb{N}^*)</math>
See also
- Fibonacci number
- Golden ratio
- Discrete cosine transform
- Discrete Fourier transform
- Wavelet transform
- List of Fourier-related transforms
- JPEG 2000
References
- Sun, Y.K.(2005). Wavelet Analysis and Its Applications. China Machine Press. ISBN 7111158768
- Jin, J.F.(2004). Visual C++ Wavelet Transform Technology and Engineering Practice. Posts & Telecommunications Press. ISBN 7115119597
- He, B.;Ma, T.Y.(2002). Visual C++ Digital Image Processing. Posts & Telecommunications Press. ISBN 7115109559
- Ingrid Daubechies(1992). Ten Lectures on Wavelets. Society for Industrial and Applied Mathematics. ISBN 0-89871-274-2