915 1995 182-191 182 Wavelet (Sumiko Hiyama) (Takao Hanada) Abstract Wavelet Wavelet Daubechies Wavelet bi-orthogonal Wavelet spiral Daubechies Wavelet 1 3 11 19 $\mathrm{b}\mathrm{i}$-orthogonal Wavelet cusp ( ) Wavelet 1 $P$ $P=(p\mathrm{o}p_{1} \ldotspn);p_{0}p_{1}$ $\ldotspn$ - 2 1 $P$ $Q$ 2 $P$ 2 1 2 $[1][2]$ $Q$ 1 $P$ : $Q$ : $\cup Q P $ $P$ $Q$ T ( $T:Parrow Q$ ) $T$ Q $Q$ 2 p: P q: Q
183 $\Psi$ $\mathrm{e}$ $P$ $Q$ $E= P-Q = \int_{p}\min_{q\epsilon Q} p-q dp=\int_{p} p-\psi(p) dp$ T Wavelet Daubechies Wavelet (DWT) Wavelet Wavelet Wavelet \langle Wavelet Daubechies Wavelet 2 FFT \S 2 2 Wavelet (MRA) Wavelet 21 MRA Wavelet MRA $L^{2}(R)$ $\{V_{j;}j\in Z\}$ 4 $\subset V_{1}\subset V_{0}\subset V_{1}$ C $cl_{os}ure_{l}2\cup=l^{2}(r)$ $\cap V_{j\in Z}=\{0\}$ $j\in Z$ $V_{j+1}=V_{j}\oplus W_{j}$ $f(x)\in V_{j}\Leftrightarrow f(2x)=\in V_{j+1}$ $\exists\varphi(x)\in V_{0}st\{\varphi(X-k);k\in Z\}$ (1) $V\mathit{0}$ $V_{j}$ $V_{j+1}$ $W_{j}$ $V_{j+1}=V_{j}\oplus W_{j}$ (2) $\{W_{j}\}$ $L^{2}(R)=\oplus W_{j}j\in Z$ (3) Wavelet \psi $W0$ $\{\psi(x-k);k\in Z\}$ $W0$ $[3][5]$ (1)(2) \mbox{\boldmath $\varphi$}(x) $\{2^{j/2}\varphi(2^{j}X-k);k\in Z\}$ $V_{0}$ $V_{j}$
$\downarrow$ 184 Wavelet $\{2^{j/2}\psi(2jx-k);k\in Z\}$ $\psi$ REJECT}$ $W_{j}$ $\ovalbox{\tt\small $V_{0}$ $\varphi$ (1) $(2)$ $\varphi(x)$ $=$ $\sum_{k}\alpha_{k}\sqrt{2}\varphi(2x-k)$ $\psi(x)$ $=$ $\sum_{k}\beta_{k}\sqrt{2}\psi(2x-k)$ $\beta_{k}$ $=$ $(-1)^{k}\alpha_{1-}k$ (4) 22 Wavelet $\{C_{m)}^{1}m\in Z\}$ $V_{1}$ $f$ $\{C_{m}^{1} ; m\in Z\}$ $f(x)$ $=$ $\sum_{m}c_{m}^{1}\sqrt{2}\varphi(2x-m)\in V_{1}$ (5) (2) $V_{1}=V0\oplus W_{0}$ $f(x)\in V_{1}$ $V_{0}$ $\{C_{l}^{0}\}$ $\{D_{l}^{0}\}$ $W_{0}$ $V0$ $W_{0}$ $f(x)= \sum_{l}c_{l}0\varphi(x-l)+\sum_{l}d_{\}}^{0}\psi(x-\iota)$ $C^{1}$ $C^{0}$ $D^{0}$ $\{\sqrt{2}\varphi(2x-k)\}$ $C_{l}^{0}$ $=$ $\int f(x)\varphi(x-l)dx=\sum_{m}\alpha m-2\iota C_{m}1$ $D_{l}^{0}$ $=$ (6) $\sum_{m}\beta_{m-2\mathfrak{s}}c^{1}m$ $(\mathrm{d}\mathrm{w}\mathrm{t})[12]$ Wavelet \alpha \beta $\beta_{k}=(-1)^{k}\alpha_{1-k}$ (7) (6) 3 Wavelet (DIWT) (6) $\{C_{l}^{0}\}$ $\{D_{l}^{0}\}$ $\{C_{m}^{1}\}$ $C_{m}^{1}= \sum_{l}\alpha_{m-2}\iota Cl0+\sum_{l}\beta_{m-}2lD_{l}0$ $\{C_{m}^{1}\}$ Wavelet $\{C_{l}^{0}\}$ $\{D_{l}^{0}\}$ $C^{0}$ 2 $D^{0}$ $C^{1}$ $\{C_{l}^{0}\}$ $C^{-1}D^{-1}$ 1 $C^{k}$ $k$ $k$ $C^{-k}$ $(k+1)$ $C$ $D$ $D^{*}$ 2 Wavelet 1 $C^{*}$ $C^{1}$ $C^{k}$ $D^{k}$ Wavelet $0$ ( 2)
185 the pyramidal hierarchy of DWT 1: The pyramidal hierarchy of $C^{k}\mathrm{a}\mathrm{n}\mathrm{d}D^{k}$ $\langle$ Wavelet Wavelet 1 P $x$ y Wavelet 2 Wavelet Wavelet Daubechies Wavelet $\{\alpha\}$ $[4 $ $i$ 3 Wavelet $m$ $C^{*}$ 4 Wavelet 1 $Q$ 5 $Q$ P $0$ (2 C $m$ D $m$ D bit ) 23 (Bi-orthogonal) Wavelet Daubechies Wavelet Wavelet Haar Wavelet $\tilde{\beta}$ \alpha \beta \alpha \tilde Bi-orthogonal Wavelet $C_{l}^{0}$ $=$ $\sum_{m}\alpha_{m-}{}_{2l}c_{m}^{1}$
$\tilde{\beta}_{k}$ 186 $-\mathrm{c}^{-v}$ 2: The dirrerent lebels of wavelet coefficients $D_{l}^{0}$ $=$ $\sum_{m}\beta_{m-}21c_{m}l$ (8) $C_{m}^{1}= \sum\tilde{\alpha}m-2\iota cll0+\sum\tilde{\beta}_{m-2}l\iota D_{\iota}^{0}$ (9) $\beta_{k}$ $=$ $(-1)^{k}\tilde{\alpha}1-k$ $=$ $(-1)^{k}\alpha 1-k$ (10) [4]pp 278-pp280 \alpha a 3 Wavelet spiral Wavelet 31 (threshold) spiral $x$ $=$ $(1/\theta)\cos\theta$ $y$ $=$ $(1/\theta)\sin\theta$
187 $02\pi\leq\theta<5\pi$ 512 \Delta \theta $=48\pi/512$ $0$ 1 Wavelet Daubechies 3 ( ) 1 $D_{3}$ 1: $\frac{ih\gamma esh_{\mathit{0}\iota dl}\gamma emanedp_{oln}is\mathrm{e}\mathrm{r}\mathrm{r}\mathrm{o}\mathrm{r}\mathrm{o}\mathrm{f}\mathrm{x}\mathrm{e}\mathrm{r}\mathrm{r}\mathrm{o}\mathrm{r}\circ \mathrm{f}\mathrm{y}}{052260115\mathrm{e}-30953\mathrm{e}-4}$ 01 52 0 $691\mathrm{E}-2$ 0 005 26 0 $459\mathrm{E}-1$ $313\bm{\mathrm{E}}- 0 $431\mathrm{E}- 2$ 2$ $\mathrm{x}$ 1 error $x$ $0$ 05 005 Wavelet Q ( 3) 05 005 A spiral applied by 3 wavelet function A spiral applied by 3 wavelet function 3: The errors of thresholds changed 32 Wavelet Daubechies Wavelet \searrow 005 05 05 2 11 3 005 11 :4
188 2: 05 Wavelet $\frac{deg_{\gamma eee}f\gamma orofxermrofy}{10276\mathrm{l}20276\mathrm{e}_{-}2}$ 3 0 $115\mathrm{L}3$ 0953E-4 11 0358E-6 0298E-6 19 0417E-6 0298 E-6 $\text{ }4$ : spiral with Gibb s phenomena and removed one 33 Daubechies Wavelet \alpha $n$ $2n-1$ Wavelet spiral $01\leq\theta<97\pi$ \Delta \theta $=97\pi/1024$ $02\pi\leq\theta<5\pi$ 11 4 4 $\overline{\dot{\tau}}-$ 225 ( 5) 15 16 256 \nu ) 225 $=05$
189 $\text{ }5$ : The original coast line consisted of 225 points 3: 05 $\frac{degree\mathrm{e}\mathrm{r}\mathrm{r}\circ \mathrm{r}\mathrm{o}\mathrm{f}\mathrm{x}\mathrm{e}\mathrm{r}\mathrm{r}\circ \mathrm{r}\mathrm{o}\mathrm{f}\mathrm{y}}{10125\mathrm{e}- 1}$ 0 $105\mathrm{k}1$ 30629E-2 11 0662E-2 $0507\mathrm{E}_{-2}$ 19 0644E-2 0625E-2 $0732\mathrm{E}-2\mathrm{E}_{- 2}$ 05 Wavelet 1 3 11 19 3 113 1 19 6 spiral 11 $=03$ spiral 001 D3 03 64 ( 7) Wavelet \S 2 Wavelet 1 Haar i $-$ Wavelet Wavelet Wavelet 4 3 0305 2 3 Wavelet \langle 2 \tau
$\bullet$ $\bullet$ by $\frac{\mathrm{o}50629\mathrm{e}_{- 207}32\mathrm{E}_{-}20723J_{-}20645\mathrm{E}- 20650\mathrm{E}-20570J- 2}{(\mathrm{c}\mathrm{a}\mathrm{s}\mathrm{e}1\cdot\cdot\circ \mathrm{r}\mathrm{d}\mathrm{e}\mathrm{r}\mathrm{d}_{\mathrm{c}}\mathrm{a}\mathrm{s}\mathrm{e}2\cdot\cdot \mathrm{i}\mathrm{n}\mathrm{v}\mathrm{e}\mathrm{r}\mathrm{s}\mathrm{e}\mathrm{o}\mathrm{r}\mathrm{d}\mathrm{e}\mathrm{r})}$ 190 Applied by degree 1 wavelet function $\mathrm{a}\mathrm{p}\mathrm{p}[\mathrm{i}\mathrm{e}\mathrm{d}$ deqree 19 wavelet function 6: The coast line reduced by $D_{1}D_{19}$ $\text{ }4$ : The errors by $\mathrm{b}\mathrm{i}$-orthogonal DWT and by orthonormal DWT $\frac{orth_{ono}rma\iota CaSe\mathit{1}orlhonormal_{C}ase\mathit{2}bi_{-}ofthogona\iota}{010129\mathrm{E}_{-}10131\mathrm{E}_{- 101}29\mathrm{E}_{-}10131\Sigma_{-}10114\mathrm{E}- 10104\mathrm{E}- 1}$ 5 Wavelet Wavelet Wavelet 13 1119 [1] D H Douglas and T K Peucker Algorithms for the Reduction of the Number of Points Required to Represent a Digitized Line or its $Ca\dot{n}catu\Gamma e$ the Canadian Cartographer $\underline{10}(1973)$ pp112-122 [2] $\underline{3}\mathrm{n}\mathrm{o}$$2(1993)$ $\mathrm{p}\mathrm{p}85_{-}104$ [3] 2 1993 [4] S Mallat Multiresolution Approximation and WaveletsTrans Amer Math $\underline{315}(1989)$ pp69-88
191 Applied by degree 3 wavelet function With $\mathrm{b}\mathrm{i}$-orthogonal DWT degree is 3 11 $\mathrm{u}$ I I $\subset-$ [ U I01 E-l $\tau=0647\mathrm{e}- \mathit{2}$ $T=\iota 9V\mathrm{O}=-\ell$ $\text{ }7$ $t$ : The coas line reduced by and $\mathrm{b}\mathrm{i}$-orthogonal wavelet $D_{3}$