Inversion formulas for the continuous wavelet transform (Q2434158)

The continuous Fourier transform of \(f\) with respect to a wavelet \(g\) is defined by \(W_gf(x,s)= \langle f, T_xD_s g \rangle\) \((x \in\mathbb R^d, s \in\mathbb R, s \neq 0)\), where \(T_x\) is the translation operator and \(D_s\) is the dilation operator. Under some conditions on \(g\) and \(\gamma\) the inversion formula holds for all \(f \in L_2(\mathbb R^d)\): \[ f = {{1}\over{C_{g,\gamma}}}\int_0^{\infty}\int_{R^d} W_gf(x,s)T_xD_s\gamma {dx ds\over s^{d+1}}, \] where the equality is understood in a vector-valued weak sense and \(C_{g,\gamma}\) is a number depending on \(g\) and \(\gamma\). In the paper under review, the author proves that under some weak conditions on \(\theta\), \(g\) and \(\gamma\), \[ f = \lim_{S\longrightarrow 0}{{1}\over{C_{g,\gamma}}} \int_0^{\infty}\int_{\mathbb R^d} \theta\left({S\over s}\right ) W_gf(x,s)T_xD_s\gamma {dx ds\over s^{d+1}} \] a.e. and in Wiener amalgam norm \(W(L_p,\ell_q)\) \((1 \leq p,q \leq \infty)\). If \(\theta = {\mathbf 1}_{[0,1]}\) then the a.e. and norm convergence hold for all \(f \in L_p(R^d)\) \((1 < p < \infty)\). The author also investigates the convergence in Pringsheim's sense and gives some examples for the summability functions \(\theta\). This inversion theorem generalizes some results of \textit{M. Wilson} [Weighted Littlewood-Paley theory and exponential-square integrability. Berlin: Springer (2008; Zbl 1138.42011)], \textit{M. Wilson} [J. Fourier Anal. Appl. 16, No. 5, 768--785 (2010; Zbl 1204.42029)], \textit{M. Rao, H. Šikić} and \textit{R. Song} [Control Cybern. 23 No. 4, 761--771 (1994; Zbl 0816.42018)], and \textit{K. Li} and \textit{W. Sun} [J. Fourier Anal. Appl. 18, No. 3, 439--455 (2012; Zbl 1250.42095)].