Signal Decomposition via Quadratic-Separable Optimization

Manuscript, December 2022.

• Manuscript

• Code

We consider the problem of decomposing a signal with missing entries into a sum of components, using the optimization formulation described by Meyers and Boyd. We propose a general form for the component loss functions, and an alternative algorithm based on quadratic-separable optimization. This yields a flexible and extensible method for formulating and solving signal decomposition problems.