NewEvery arXiv paper, its researchers & institutions — mapped.
mathematics

$αβ$-Tessarine Toolbox: A high-performance MATLAB framework for hypercomplex tensor algebra

arXiv:2607.27809

summary

The paper presents the αβ‑Tessarine Toolbox, a high‑performance MATLAB framework for performing hypercomplex tensor algebra, including exact matrix factorizations, aimed at large‑scale image processing and machine‑learning tasks.

Abstract

The \textit{$αβ$-Tessarine Toolbox} is introduced as a high-performance MATLAB framework for generalized hypercomplex tensor algebra. By implementing the $αβ$-tessarine mathematical structure, a strictly more general environment than existing quaternion or limited tessarine packages is provided. This open-source software bridges abstract theoretical models and experimental applications, allowing researchers to efficiently solve multidimensional problems. Built-in capabilities include exact hypercomplex matrix factorizations, such as Singular Value Decomposition (SVD) and QR, optimized for large-scale image in-painting, denoising, and digital watermarking. The scalable architecture facilitates new computational research in machine learning and multidimensional signal processing.

Topics & keywords

#hypercomplex algebra#tensor computations#MATLAB toolbox#image processing#machine learningαβ‑tessarinehypercomplex tensor algebrasingular value decompositionQR factorizationimage in‑paintingdigital watermarking