Bendlets: a second-order shearlet transform with bent elements
Extending the idea of shearlets, this paper further studies the ``bending operations and defines a new shearlet-like system, called \textit{bendlets}. Such systems can then be used to classify functions with 2D discontinuity up to its curvature in the asymptotic sense. \par Consider an $\alpha$-scaling operator $A_{a,\alpha}$, extending the shearing operator $S_(s,b)$ (second order) as follows: \[ A_{a,\alpha}:=\begin{pmatrix}a & 0\\0 & a^\alpha\end{pmatrix}, S_{(s,b)}\begin{pmatrix}x_1\\x_2\end{pmatrix}:=\begin{pmatrix}1 & s\\0 & 1\end{pmatrix}\begin{pmatrix}x_1\\x_2\end{pmatrix}+\begin{pmatrix}1 & bx_2\\0 & 1\end{pmatrix}\begin{pmatrix}x_1\\x_2\end{pmatrix}, \] where the parameter $s$ is with respect to the classical shearing parameter while $b$ corresponds to the bending parameter. A \textit{bendlet} system can be defined as \[ BS_{\psi}^\alpha:=\{\psi_{a,s,b,t} | (a,b,s,t)\in \mathbb{R}^+\times \mathbb{R}\times \mathbb{R} \times \mathbb{R}^2\}, \] where $\psi\in L_2(\mathbb{R}^2)$ and \[ \psi_{a,b,s,t}:=a^{-(1+\alpha)/2}\psi(A_{a,\alpha}^{-1}S_{(-s,-b)}(\cdot-t)). \] Then \textit{bendlet transform} associated with the bendlet system $BS_\psi^\alpha$ for a function $f\in L_2(\mathbb{R}^2)$ is given by \[ [\mathcal{BS}_\psi^\alpha (f)](a,s,b,t):=\langle f, \psi_{a,s,b,t} \rangle. \] To avoid extreme shearing and bending that results in numerical infeasibility, employing the cone-adaptive approach similar to the cone-adaptive shearlets, the paper defines the \textit{cone-adaptive bendlet system} by splitting $\mathbb{R}^2$ into horizontal and vertical cones: \[ BS_{\psi}^\alpha:=\{\psi_{a,s,b,t,\iota} | a\in (0,1), s\in[-1,1], b\in \mathbb{R}, t\in \mathbb{R}^2, \iota\in\{-1,1\}\}, \] where $\iota\in\{-1,1\}$ indicates the horizontal or vertical cone, respectively. \par Consider $\psi = \psi^1\otimes\phi^1$ with $\psi^1$ a 1D compactly supported wavelet having $M$-th order vanishing moments and $\phi^1\in C^L(\mathbb{R})$ for some $L>M$. The paper then presents the main results regarding the characterizations of the singularities of a simple function $\chi_D\in L_2(\mathbb{R}^2)$ with $D\subset\mathbb{R}^2$ and $\partial D\in C^\infty(\mathbb{R}^2)$, using the cone-adaptive bendlet systems $BS_\psi^\alpha$. The paper shows that: \par (a) If $p\notin\partial D$, then the coefficient $|\langle \chi_D, \psi_{a,s,b,p,\iota}\rangle| \lesssim a^N$ for all $a\in(0,1), s\in[-1,1], b\in\mathbb{R}, \iota\in\{-1,1\}$, and $N\in\mathbb{N}$. \par (b) If $p\in\partial D$ but $\psi_{a,s,b,p,\iota}$ is not aligned with respect to the normal of $\partial D$ at $p$, then $|\langle \chi_D, \psi_{a,s,b,p,\iota}\rangle| \lesssim a^{(1-\alpha)(M+1)+\frac{1+\alpha}{2}}$. \par (c) If $p\in\partial D$, $\psi_{a,s,b,p,\iota}$ is aligned with respect to the normal of $\partial D$ at $p$ but not to the curvature $b$ of $\partial D$ at $p$, then $|\langle \chi_D, \psi_{a,s,b,p,\iota}\rangle| \lesssim a^{\frac{2-\alpha}{2}}$. \par (d) If $p\in\partial D$ and $\psi_{a,s,b,p,\iota}$ are not aligned with respect to the normal and the curvature of $\partial D$ at $p$, then $|\langle \chi_D, \psi_{a,s,b,p,\iota}\rangle| \lesssim a^{\frac{1+\alpha}{2}}$. \par The last part of the paper is numerical implementation, demonstration, and further discussion.
- Bendlet transforms: a mathematical perspective
- Abstract shearlet transform
- The continuous shearlet transform in arbitrary space dimensions
- Shearlet transforms and directional regularities
- Nonseparable Shearlet Transform
- The Discrete Shearlet Transform: A New Directional Transform and Compactly Supported Shearlet Frames
- scientific article; zbMATH DE number 7047359
- Shearlets: theory and applications
- On the shearlet transform using hyperbolic functions
- Shearlet coorbit spaces, shearlet transforms and applications in imaging
- \(\alpha\)-Molecules
- A Shearlet Approach to Edge Analysis and Detection
- Analysis and detection of surface discontinuities using the 3D continuous shearlet transform
- Cartoon approximation with \(\alpha\)-curvelets
- Characterization and analysis of edges in piecewise smooth functions
- Characterization and Analysis of Edges Using the Continuous Shearlet Transform
- Classification of edges using compactly supported shearlets
- Compactly supported shearlets are optimally sparse
- Construction of compactly supported shearlet frames
- Continuous curvelet transform. I: Resolution of the wavefront set
- Continuous curvelet transform. II: Discretization and frames
- Continuous frames, function spaces, and the discretization problem
- Continuous shearlet frames and resolution of the wavefront set
- Detection of boundary curves on the piecewise smooth boundary surface of three dimensional solids
- Edge analysis and identification using the continuous shearlet transform
- Filtering, segmentation and depth
- New tight frames of curvelets and optimal representations of objects with piecewise C2 singularities
- Optimally sparse approximations of 3D functions by compactly supported shearlet frames
- Optimally Sparse Multidimensional Representation Using Shearlets
- Resolution of the wavefront set using continuous shearlets
- Resolution of the wavefront set using general continuous wavelet transforms
- ShearLab 3D: faithful digital shearlet transforms based on compactly supported shearlets
- Shearlet coorbit spaces and associated Banach frames
- Shearlets. Multiscale analysis for multivariate data.
- Sparse components of images and optimal atomic decompositions
- Sparse multidimensional representations using anisotropic dilation and shear operators
- THE UNCERTAINTY PRINCIPLE ASSOCIATED WITH THE CONTINUOUS SHEARLET TRANSFORM
- Retinal blood vessel segmentation from fundus image using an efficient multiscale directional representation technique bendlets
- Higher order analysis of the geometry of singularities using the Taylorlet transform
- Mini-workshop: Interpolation, approximation, and algebra. Abstracts from the mini-workshop held February 13--19, 2022
- Bendlet transforms: a mathematical perspective
- Efficient representation of spatio-temporal data using cylindrical shearlets
- Registration and inpainting of biological slice images based on Bendlet transform and interval Shannon-cosine wavelet
- Shearlets: from theory to deep learning
- Tighter uncertainty principles associated with bendlet and quaternion-Bendlet transforms
This page was built for publication: Bendlets: a second-order shearlet transform with bent elements
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1713642)