Year: 2024
Type: article
Abstract: We store and exchange more digital images than ever before. Image qualities are often reduced to decrease the cost of storage and transfer of the images. Digital image compression is a technique that reduces the size of an image while preserving its quality to be acceptable for a particular purpose. Matrix factorizations are widely used for this purpose. This paper presents an application of the QKRK and block QKRK factorizations of a matrix to image compression. We have conducted a series of experiments using Matlab software. This paper presents the comparative analysis of the compressed images using the QR, SV D, QKRK, and block QKRK factorizations. The similarity between the original and compressed images is measured using the L2-norm and structural similarity. It is demonstrated that using the QKRK factorization for image compression allows the achievement of the desired quality of a fragment of the image compared to the rest of the image and is also computationally efficient.
Authors: Erik Eckenberg, Knarik Tunyan
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W4406779283