20 Sep The emergency of the Singular Value Decomposition (SVD) is considered a new paradigm that helps in the processing of various types of images. Despite the wide utilization of the SVD&ap

Lab

CS 535 Numerical Computation

Name: Yongseung Lee

Coyote ID: 004174533

I Problem Statement and Goals

The emergency of the Singular Value Decomposition (SVD) is considered a new paradigm that helps in the processing of various types of images. Despite the wide utilization of the SVD's generous properties, there are various properties of the SVD that remains unexploited as far as image processing is concerned. The paper seeks to examine these unexploited in addition to the introduction of the new trends as well as challenges while utilizing the SVD in image processing applications.

II Methods, Algorithms, Tools used (if any)

The aim is only to examine the various properties of the SVD Image. Comprehension about the significance/ importance of the SVD, the SVD is quite a powerful and reliable orthogonal matrix decomposition method.

III Program (provide detailed codes for each question if any)

No codes were used in this research.

IV Results (provide analysis and results for each question)

This involved calculating the comprehension ratio, which in this case is calculated using the following formula.

R= *10

,

Applications of the SVD

Yongseung Lee (004174533)

School of Computer Science and Engineering, CSUSB

Abstract

The emergency of the Singular Value Decomposition (SVD) is considered a new paradigm that helps in the processing of various types of images. The SVD is currently one of the attractive algebraic transforms for image processing applications. This paper aims to propose an experimental survey for the SVD as an efficient transform in image processing applications.

Even though it is known that the SVD provides an attractive property as far as the imaging is concerned, the exploration through the utilization of the various properties in different image applications has not developed that much, but it is still developing.

Generally, because the SVD has various striking properties that haven’t been used, the paper provides significant contribution through the utilization of the generous properties in the new image applications while at the same time providing good recommendations that can be adopted when carrying out more research challenges.

Therefore, the various properties for the SVD for the images are presented experimentally to be used in the development of the new SVD-based image processing applications. The article provides a survey regarding the developed SVD based image applications. In addition to that, the paper recommends or proposes various new contributions by the SVD properties analysis in various image processing.

In general, this paper seeks to offer an efficient as well as a better comprehension regarding the SVD in the image processing while at the same time identifying the essential applications as well as the open research directions in the area that is typically considered significant; the SVD based image processing in the future research.

1. Introduction

The SVD refers to the optimal matrix decomposition in the least square sense, which packs the extreme and maximum signal energy into as few coefficients as possible. The Singular value decomposition (SVD) is considered a stable and effective method of splitting the system into a set of linearly independent components; every component bears its energy contribution.

The Singular value decomposition (SVD) refers to the numerical technique that helps diagonalize the matrices in the numerical analysis. In this regard, the SVD is therefore considered an attractive algebraic transform for image processing. The main reason for this consideration is because the SVD has numerous advantages. For instance, it has a maximum energy packing that is normally utilized in the compression, and it can manipulate the images in the base of the two distinctive subspaces data as well as noise subspaces that are normally utilized in what is called the noise filtering as well as in the watermarking applications. In this regard, each application focuses on exploiting the main properties of the SVD.

Moreover, it is further utilized in solving the least square problems, in the computation of the pseudo-inverse of the matrix and the multivariate analysis. The SVD is a strong as well as reliable orthogonal matrix decomposition method that is generally based on its conceptual as well as stability reasons that help to become quite popular in the processing of the signal. The SVD has the capability of adapting the variations in the local statistics of the image. This article aims to provide thorough experiments for the great properties of the SVD, which have not been fully examined in digital image processing. Additionally, the focus will also be on the developed SVD based on the image processing methods that focused on the compression, watermarking, and quality measures. Therefore, the experiments focus on validating the various known but the SVD’s properties that are not used in image processing applications. Therefore, the paper contributes to utilizing the SVD generous properties, which are generally unexploited in image processing.

Additionally, the paper provides an introduction to the new trends and challenges while utilizing the SVD in image processing applications. The examination of the various new trends is done experimentally, which is then validated while various are demonstrated through requires more work for the full validation. The paper, therefore, allows the various tracks for the future work for the utilization of the SVD considered to be an imperative tool for signal processing.

2. Related work

II. SINGULAR VALUE DECOMPOSITION (SVD)

As far as linear algebra is concerned, the SVD refers to the factorization of the rectangular real as well as complex matrix analogous to the diagonalization of the symmetric as well as the Hermitian square matrices through the use of the basis of eigenvectors. The SVD is considered stable and an effective method of splitting the system into a set of linearly independent components, such that every component bears an individual energy contribution [1, 3]. In this case, a digital Image X of the size MxN, with M≥N, may be represented by its SVD as follows;

In this case, You refers to an MxM orthogonal matrix, V refers to an NxN orthogonal matrix. In contrast, S refers to an MxN matrix having the diagonal elements representing the singular values, si of the X [8].

The utilization of the subscript T is meant to denote the transpose of the matrix. Additionally, the columns of the orthogonal matrix usually are referred to as the left singular vectors. In contrast, the columns of the orthogonal matrix V are commonly referred to as the right singular vectors [12]. The left singular vectors (LSCs) of the X are generally the eigenvectors of the XXT, while the right singular vectors (RSCs) of the X are generally called the eigenvectors of the XTX. Generally, every singular value (SV) helps in specifying the luminance of the image layer, and the pair of singular vectors (SCs) helps in specifying the geometry of the image.

Generally, U and V are considered unitary orthogonal matrices, while on the other hand, S is considered a diagonal matrix of reducing singular values. Generally, the singular value of every Eigen image may be determined by just identifying the 2-norm [5].

In this regard, because SVD focuses on maximizing the largest singular values, usually, the initial Eigen image is considered the pattern which accounts for the greatest amount of the variance-covariance structure. The

3. Method

In this case, there isn’t much to write about the method. However, the aim is to examine the various properties of the SVD Image. Generally, it is first of all important to understand that SVD is typically a strong as well as reliable orthogonal matrix decomposition method. In this regard, as a result of the stability of the SVD, this is why it has become more and more popular as far as signal processing is concerned. Generally, the SVD has become an attractive algebraic transform for image processing.

The SVD incorporates significant properties as far as imaging is concerned. The various properties of the SVD are usually advantageous for the images; some of these characteristics include the maximum energy packing, its capability of solving the least-squares problems, the capability of computing the pseudo-inverse of a particular matrix as well, as the capability of achieving a multivariate analysis [1,2].

Generally, the basic property includes its relation to the matrix rank as well as its capability of being able to approximate the matrices of a specifically provided rank. The low-rank matrices are typically utilized in the representation of the digital images, and this facilitates the description through the utilization of the sum of quite a small set of the Eigen images [3,4]. In this case, it is therefore essential to focus on proving some of the hypotheses.

4. Experiment and dataset

In this case, there isn’t much to write about the method. However, the aim is to examine the various properties of the SVD Image. Generally, it is first of all important to understand that SVD is typically a strong as well as reliable orthogonal matrix decomposition method. In this regard, as a result of the stability of the SVD, this is why it has become more and more popular as far as signal processing is concerned. Generally, the SVD has become an attractive algebraic transform for image processing.

The SVD incorporates significant properties as far as imaging is concerned. The various properties of the SVD are usually advantageous for the images; some of these characteristics include the maximum energy packing, its capability of solving the least-squares problems, the capability of computing the pseudo-inverse of a particular matrix as well, as the capability of achieving a multivariate analysis [1,2].

Generally, the basic property includes its relation to the matrix rank as well as its capability of being able to approximate the matrices of a specifically provided rank. The low-rank matrices are typically utilized in the representation of the digital images, and this facilitates the description through the utilization

5. Timeline

In this section, a project plan should be described.

References

[1] M Moonen, P van Dooren, J Vandewalle, “Singular value decomposition updating algorithm for subspace tracking," SIAM Journal

on Matrix Analysis and Applications (1992)

[2] T. Konda, Y. Nakamura, A new algorithm for singular value decomposition and parallelization, Parallel Compute. (2009), DOI: 10.1016/j.parco.2009.02.001

[3] H. C. Andrews and C. L. Patterson, “Singular value decompositions and digital image processing,” IEEE Trans. on Acoustics, Speech, and Signal Processing, vol. ASSP-24, pp. 26–53, 1976.

[4] Julie L. Kamm, “SVD-Based Methods for Signal and Image

Restoration", Ph.D. Thesis (1998)

[5] J.F. Yang and C.L. Lu,” Combined Techniques of Singular Value Decomposition and Vector Quantization for Image Coding,” IEEE Trans. Image Processing, pp. 1141 – 1146, Aug. 1995.

[6] Xiaowei Xu, Scott D. Dexter, Ahmet M. Eskicioglu: A hybrid scheme for encryption and watermarking. Security, Steganography, and Watermarking of Multimedia Contents 2004:725-736

[7] K. Konstantinides, B. Natarajan, and G.S. Yovanof,” Noise Estimation and Filtering Using Block-Based Singular Value Decomposition,” IEEE Trans. Image Processing, vol. 6, pp. 479- 483, March 1997.

[8] E. Ganic and A. M. Eskiciogulu, Secure DWT-SVD Domain Image

Watermarking: Embedding Data in All Frequencies, ACM Multimedia and Security Workshop 2004, Magdeburg, Germany, September 20-21, 2004

[9] V.I. Gorodetski, L.J. Popyack, V. Samoilov, and V.A. Skormin,” SVD- Based Approach to Transparent Embedding Data into Digital Images," Proc. Int. Workshop on Mathematical Methods, Models and Architecture

,