Splines are Universal Solutions of Linear Inverse Problems with Generalized-TV regularization

Abstract

Splines come in a variety of flavors that can be characterized in terms of some differential operator L. The simplest piecewise-constant model corresponds to the derivative operator. Likewise, one can extend the traditional notion of total variation by considering more general operators than the derivative. This leads us to the definition of the generalized Beppo-Levi space M, which is further identified as the direct sum of two Banach spaces. We then prove that the minimization of the generalized total variation (gTV) over M, subject to some arbitrary (convex) consistency constraints on the linear measurements of the signal, admits nonuniform L-spline solutions with fewer knots than the number of measurements. This shows that non-uniform splines are universal solutions of continuous-domain linear inverse problems with LASSO, L1, or TV-like regularization constraints. Remarkably, the spline-type is fully determined by the choice of L and does not depend on the actual nature of the measurements.

28 pages, 1 figure

References in corpus (1)

• Sparsity and Incoherence in Compressive Sampling

Cited by in corpus (23)

• The Role of Neural Network Activation Functions
• Sparsity of solutions for variational inverse problems with finite-dimensional data
• Continuous-Domain Solutions of Linear Inverse Problems with Tikhonov vs. Generalized TV Regularization
• What Kinds of Functions do Deep Neural Networks Learn? Insights from Variational Spline Theory
• Near-Minimax Optimal Estimation With Shallow ReLU Neural Networks
• Deep Neural Networks with Trainable Activations and Controlled Lipschitz Constant
• Deep Learning Meets Sparse Regularization: A Signal Processing Perspective
• A generalized conditional gradient method for dynamic inverse problems with optimal transport regularization
• Sampling and Reconstruction of Sparse Signals in Shift-Invariant Spaces: Generalized Shannon's Theorem Meets Compressive Sensing
• On the extremal points of the ball of the Benamou-Brenier energy
• TV-based Reconstruction of Periodic Functions
• A superposition principle for the inhomogeneous continuity equation with Hellinger-Kantorovich-regular coefficients
• Periodic Splines and Gaussian Processes for the Resolution of Linear Inverse Problems
• High-Contrast Reflection Tomography with Total-Variation Constraints
• Functional Penalised Basis Pursuit on Spheres
• A Fast and Scalable Polyatomic Frank-Wolfe Algorithm for the LASSO
• Extremal points and sparse optimization for generalized Kantorovich-Rubinstein norms
• On the Uniqueness of Solutions for the Basis Pursuit in the Continuum
• Controlled Learning of Pointwise Nonlinearities in Neural-Network-Like Architectures
• Single-Pixel Compressive Imaging in Shift-Invariant Spaces via Exact Wavelet Frames
• Function-Space Optimality of Neural Architectures with Multivariate Nonlinearities
• RINS-T: Robust Implicit Neural Solvers for Time Series Linear Inverse Problems
• On extreme points and representer theorems for the Lipschitz unit ball on finite metric spaces