A Least-Squares Problem of a Linear Tensor Equation of Third-Order for Audio and Color Image Processing
Pablo Soto-Quirós · 2022
This paper describes a new framework for solving inverse tensor problems for denoising data. This inverse tensor problem can be represented as a least-squares problem of a linear tensor equation of third-order. We use training data to replace knowledge of the noise model by introducing a framework that avoids including a particular noise model. Additionally, we describe a numerical method based on a tensor pseudoinverse to estimate a solution to the least-squares problem. Finally, we present two real-life applications for audio denoising and color image deconvolution. The advantages of the considered technique are discussed and illustrated numerically.