Modified Levenberg-Marquardt Algorithm For Tensor CP Decomposition in Image Compression

This paper explores a new version of the Levenberg-Marquardt algorithm usedfor Tensor Canonical Polyadic (CP) decomposition with an emphasis on imagecompression and reconstruction. Tensor computation, especially CPdecomposition, holds significant applications in data compression and analysis.In this study, we formulate CP as a nonlinear least squares optimizationproblem. Then, we present an iterative Levenberg-Marquardt (LM) based algorithmfor computing the CP decomposition. Ultimately, we test the algorithm onvarious datasets, including randomly generated tensors and RGB images. Theproposed method proves to be both efficient and effective, offering a reducedcomputational burden when compared to the traditional Levenberg-Marquardttechnique.