Asymptotic performance of projection estimators in standard and hyperbolic wavelet bases

Abstract

International audienceWe provide a novel treatment of the ability of the standard (wavelet-tensor) and of the hyperbolic (tensor product) wavelet bases to build nonparametric estimators of multivariate functions. First, we give new results about the limitations of wavelet estimators based on the standard wavelet basis regarding their inability to optimally reconstruct functions with anisotropic smoothness. Next, we provide optimal or near optimal rates at which both linear and non-linear hyperbolic wavelet estimators are well-suited to reconstruct functions from anisotropic Besov spaces and subsequently we characterize the set of all the functions that are well reconstructed by these methods with respect to these rates. As a first main result, we furnish novel arguments to understand the primordial role of sparsity and thresholding in multivariate contexts, in particular by showing a stronger exposure of linear methods to the curse of dimensionality. Second, we propose an adaptation of the well known block thresholding method to a hyperbolic wavelet basis and show its ability to estimate functions withanisotropic smoothness at the optimal minimax rate. Therefore, we prove the pertinence of horizontal information pooling even in high dimensional settings. Numerical experiments illustrate the finite samples properties of the studied estimators

Hyperbolic Wavelet-Fisz Denoising for a Model Arising in Ultrasound Imaging

“…Hyperbolic wavelet bases are unconditional bases for functions in L 2 ([0, 1] 2 ). They provide sparse representations so that the simple hard thresholding procedure which consists in keeping only coefficients with magnitude larger than a given threshold; setting the others to zero, provide estimators with very good theoretical and practical performances [7] [8].…”

“…In fact, this image is an additive model when the regularities in the two space dimensions are independent. This is a highly anisotropic case where the hyperbolic setting gives optimal results [8]. The second Kidney image is a CT image taken from FIELD II's website 3 .…”

“…We extend this method to hyperbolic wavelets and show how variance stabilization can be easily performed using the low frequency outputs from the wavelet transform at different scales. The motivation behind the use of hyperbolic wavelets is their capacity to provide better estimators than the standard wavelet-tensor construction whenever images contain anisotropic features [7] [8]. This situation often occurs in ultrasound imaging due to the presence of features skin layers, vessels.…”

On Mixed Fractional Lifting Oscillation Spaces

“…These developments together with the interrelations with hyperbolic crosses have numerous applications in computational mathematics, the numerical solution of partial differential equations, data analysis and signal processing [17,23,24,[30][31][32][33][34]. In particular, ref.…”

Multivariate intensity estimation via hyperbolic wavelet selection

“…These indeed have the advantage of being as easily tractable as their univariate counterparts since each isotropic wavelet is a tensor product of univariate wavelets coming from the same resolution level. Notable counterexamples are [Don97], [Neu00] and [NvS97], or [ACF15] and [ACF14]. They underline the usefulness of hyperbolic wavelet bases, where coordinatewise varying resolution levels are allowed, so as to recover a wider range of functions, and in particular functions with anisotropic smoothness.…”