Directional polynomial wavelets on spheres
The article develops a mathematically rigorous framework for constructing discrete tight frames on the sphere \(S^{d-1}\), with particular emphasis on directional polynomial wavelets. The work extends classical isotropic constructions by introducing adjustable directionality, enabling the analysis of anisotropic structures such as edges or oriented features. This represents an important step beyond zonal wavelet systems, which depend only on geodesic distance and therefore lack orientation sensitivity.\N\NA central contribution is the construction of wavelets whose continuous transform naturally lives on the rotation group \(SO(d)\), rather than solely on the sphere. While this increases complexity, the author shows that limiting directionality via a parameter \(K\) significantly reduces the required sampling density. The resulting systems remain computationally feasible while retaining directional sensitivity. The tight frame property is established through Parseval-type identities, e.g.\N\[\N\sum_{j,\ell,m} |\langle f, \psi^{K}_{j,\ell,m}\rangle|^2 = \|f\|_{L^2(S^{d-1})}^2,\N\]\Nwhich guarantees stable reconstruction and energy preservation.\N\NThe mathematical development relies heavily on harmonic analysis, including spherical harmonics \(Y^{d,n}_k\), Gegenbauer polynomials, and representation theory of \(SO(d)\). Localization estimates demonstrate strong spatial decay:\N\[\N|\psi^K_N(\eta)| \le C N^{d-1}(1+N\,\mathrm{dist}(\eta,e_d))^{-q},\N\]\Nwhich is crucial for approximation efficiency in \(L^p\)-spaces. The paper also provides convergence results and discusses steerability, making the framework attractive for applications in signal processing or spherical data analysis.\N\NOne strength of the article is its systematic generalization from \(S^2\) to higher dimensions, maintaining clear connections to earlier literature while introducing new theoretical insights. However, the technical depth and heavy notation may be demanding for readers unfamiliar with representation theory or spherical harmonic analysis.\N\NOverall, the article presents a significant advancement in directional wavelet theory on manifolds, offering both theoretical depth and potential computational relevance.
- Approximation Theory and Harmonic Analysis on Spheres and Balls
- Characterizations of function spaces on the sphere using frames
- Decomposition of Besov and Triebel-Lizorkin spaces on the sphere
- Direction sensitive analysis of higher order jump discontinuities along circles on the sphere
- Flexible-bandwidth needlets
- scientific article; zbMATH DE number 1531949 (Why is no real title available?)
- scientific article; zbMATH DE number 3274385 (Why is no real title available?)
- Localisation of directional scale-discretised wavelets on the sphere
- Localized polynomial frames on the interval with Jacobi weights
- Localized Tight Frames on Spheres
- On the representation of smooth functions on the sphere using finitely many bits
- Polynomial frames on the sphere
- Polynomial wavelet-type expansions of functions on a sphere.
- Second-Generation Curvelets on the Sphere
- Spherical Marcinkiewicz-Zygmund inequalities and positive quadrature
This page was built for publication: Directional polynomial wavelets on spheres
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6867298)
