Enhancing the SVD Compression

TL;DR

This paper introduces E-SVD, a lossless enhancement of traditional SVD that eliminates orthonormality constraints, enabling more efficient data compression especially at high fidelity levels, with theoretical proof and empirical validation.

Contribution

The paper provides a theoretical proof that orthonormality constraints in SVD can be fully eliminated losslessly, and develops E-SVD, a novel method that improves compression efficiency.

Findings

E-SVD reduces storage by 25% compared to SVD at high fidelity.

Theoretical proof of lossless elimination of orthonormality constraints.

Empirical validation in remote sensing and IoT scenarios confirms E-SVD's superiority.

Abstract

Orthonormality is the foundation of matrix decomposition. For example, Singular Value Decomposition (SVD) implements the compression by factoring a matrix with orthonormal parts and is pervasively utilized in various fields. Orthonormality, however, inherently includes constraints that would induce redundant degrees of freedom, preventing SVD from deeper compression and even making it frustrated as the data fidelity is strictly required. In this paper, we theoretically prove that these redundancies resulted by orthonormality can be completely eliminated in a lossless manner. An enhanced version of SVD, namely E-SVD, is accordingly established to losslessly and quickly release constraints and recover the orthonormal parts in SVD by avoiding multiple matrix multiplications. According to the theory, advantages of E-SVD over SVD become increasingly evident with the rising requirement of data fidelity. In particular, E-SVD will reduce 25% storage units as SVD reaches its limitation and fails to compress data. Empirical evidences from typical scenarios of remote sensing and Internet of things further justify our theory and consistently demonstrate the superiority of E-SVD in compression. The presented theory sheds insightful lights on the constraint solution in orthonormal matrices and E-SVD, guaranteed by which will profoundly enhance the SVD-based compression in the context of explosive growth in both data acquisition and fidelity levels.