Scribd
Upload
147 views5 pages

Chapter 1: Introduction: 1.1 Background and Motivation

This chapter introduces the background and motivation for developing algorithms to reduce noise in low-dose CT images. It discusses how CT imaging works and the tradeoff between image quality and radiation dose. Developing efficient denoising algorithms is important to reduce noise in low-dose CT images while maintaining diagnostic quality. Conventional denoising methods are not well-suited for low-dose CT images. Wavelet-based methods show promise for effective low-dose CT denoising. The remainder of the thesis will analyze wavelet transform methods for low-dose CT denoising and compare to conventional methods.

Uploaded by

nirav34
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
147 views5 pages

Chapter 1: Introduction: 1.1 Background and Motivation

This chapter introduces the background and motivation for developing algorithms to reduce noise in low-dose CT images. It discusses how CT imaging works and the tradeoff between image quality and radiation dose. Developing efficient denoising algorithms is important to reduce noise in low-dose CT images while maintaining diagnostic quality. Conventional denoising methods are not well-suited for low-dose CT images. Wavelet-based methods show promise for effective low-dose CT denoising. The remainder of the thesis will analyze wavelet transform methods for low-dose CT denoising and compare to conventional methods.

Uploaded by

nirav34
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 5

Chapter 1

Introduction

Chapter 1: Introduction

1.1 Background and Motivation


The science of medical imaging owes much of its existence to the discovery of x-rays by W. C. Roentgen over 115 years ago in 1895. However, it was the development of practical computed tomography (CT) scanners in the early 1970s by G. Hounsfield and others that brought computers into medical imaging and clinical practice. Since then, computers have become integral components of modern medical imaging systems and hospitals, performing a variety of tasks from data acquisition and image generation to image display and analysis. With the widespread acceptance of computed tomography came an implicit invitation to apply computers and computing to a host of other medical imaging situations.

The increase in the number of modalities of medical imaging and in their practical use has been accompanied by an almost natural increase in the scope and complexity of the associated problems, requiring further advanced techniques for their solution. For example, CT imaging comes with lots of problems such as various types of noises and artifacts, effects of attenuation along the path of propagation of the x-rays through the body and severe blurring. Another most important issue is the level of radiation dose in CT imaging. Radiation dose has direct impact on the visual quality of the image. Higher radiation doses provide good resolution image. But the problem with high-dose CT imaging is that it creates risk of cancer to the patient. Oppositely, low radiation dose suffers from various types of noises such as quantum (random) noise which results in degraded poor quality CT images. Thus, it is highly necessary to reduce the level of noise from low-dose CT images using efficient image denoising algorithms.

The interpretation of a CT image by an expert bears the weight of the experience and expertise of the analyst, however, such analysis is almost always subjective. Denoising of low-dose CT image noises, if performed with the appropriate and efficient algorithms, has the potential to add objective strength to the interpretation of the expert. It thus becomes possible to improve the diagnostic confidence and accuracy of even an expert with many years of experience.

Chapter 1

Introduction

Developing an algorithm for CT image analysis, however, is not an easy task, quite often; it might not even be a straightforward process. Benign diseases often mimic the features of malignant diseases; malignancies may exhibit characteristic patterns which, however, are not always guaranteed to appear. Handling all of the possibilities and the degrees of freedom in a biomedical system is a major challenge in most applications. Techniques proven to work well with a certain system or set of images may not work in another seemingly similar situation.

1.2 CT Imaging Principle and Radiation Issues [1]


A CT (computed tomography) scanner is a special kind of X-ray machine. Instead of sending out a single X-ray through human body as with ordinary X-rays, CT uses several beams simultaneously from different angles to scan some certain thickness on the level of the human body. When x-ray goes through a certain part, the different densities of tissue absorb different dose of x-ray. The higher the tissue density is, such as the bones, the more the x-ray is absorbed. The detectors receive weak signal. The x-ray through the level, received by the detectors, is changed into electrical signal, and then is changed to the digital data by analog / digital converter. Finally, the data is transmitted into the computer. The data information obtained each x-ray attenuation coefficient or absorption coefficient of the voxel by calculating, and then to be re-arranged matrix, that is, digital matrix. The data in the digital matrix can be changed into small boxes ranging from black to white gray through digital/analog converter. That is, pixel, and then to be arranged by Matrix. It is called a CT image. Obviously, CT images are reconstructed images. Since 1970s, Computed Tomography (CT) technology has become important diagnostic tool and has been widely used in medical field.

An important issue within radiology today is how to reduce the radiation dose during CT examinations without compromising the image quality. Higher radiation dose limits its further application for mass screening such as the examination of people under higher risk of lung cancer, and that of children and pregnant women. Moreover, radiation dose has a direct impact on image quality due to quantum statistics. On the other hand, the low dose CT image has electronic and quantum noise that is non stationary and directional. The CT image has streak artifacts, aliasing artifacts and beam hardening effects due to improper projection on detector. International Commission on Radiological Protection (ICRP) proposed that the principles of x-ray examination should be optimized and reasonable.
2

Chapter 1

Introduction

1.3 Advantages of Low-Dose CT Imaging [1]


Thus, low-dose computed tomography (CT) is clinically desirable, especially for the screening purpose. Naidich [5] first proposed the concept of low-dose CT in 1990. A simple and cost-effective solution among many other solutions to achieve low-dose CT is to lower the x-ray tube current (mA) or deliver less x-rays (while keeping other scanning parameters unchanged).

Some of the advantages of the low-dose CT can be listed as follow: Greatly reduces the patients of the x-ray radiation dose. Used for large-scale population especially among pregnant women and children. Used for high-risk group such as lung cancer, tuberculosis and pneumoconiosis. Significantly reduces the costs of patients. Reduces the damage to CT scanners x-ray tube and detectors which can extend the life of x-ray tube.

1.4 Degradation Due to Low-Dose CT Imaging and Solutions


In 1995, Mayo [1] shown that when x-ray tube current lowers to 80 mAs, the image quality is also degraded due to the excessive x-ray quantum noise. Thus, during the process of low-dose CT imaging, the photon detectors acquiring less data will create some visual noise such as spots and snow-like structure in the reconstruction of noise. Minor noise can affect the image on the diagnosis of small lesions, while excessive noise can be serious enough to lead to misdiagnosis. Thus, it is highly necessary to reduce the noise from the low-dose CT images using various denoising algorithms.

Most of the conventional denoising algorithms, such as spatial filtering and Fourier-based spectral filtering techniques are not efficient enough for low-dose CT imaging. For example, the mean filter and Gaussian filter have the disadvantage of blurring the edges when reducing noise
[4]

. While for the median filter, it can preserve edges, but the fine structures are
[4]

suppressed and it tends to produce regions of constant or nearly constant intensity in homogeneous image regions . The adaptive minimum mean-square error filter+

outperforms the two kinds of filters mentioned above by analyzing the local image intensity statistics. However, there is no guarantee that a denoised image with a higher SNR value achieved in this way has acceptable visual quality. As demonstrated by Umbaugh [4], some
3

Chapter 1

Introduction

obvious visual artifacts may be seen in the denoised image. Fourier-based low-pass spectral filtering approaches can significantly reduce noise by suppressing the high frequency information, where noise is typically dominant. Nevertheless, these techniques often reduce the sharpness of edges due to the fact that high frequencies are also associated with edges. In recent years, wavelet transform (WT) has been widely applied to image denoising and a lot of efficient wavelet-based denoising algorithms have been developed.

Multiscale wavelet based denoising algorithms can significantly reduce the required radiation dose. The algorithm works as a filter that reduces random noise and enhances structures. Here, the approach is to eliminate the effects using wavelets. In this way, it is possible to get high-quality images and at the same time lower the dose by as much as 30 to 70 percent. Thus high quality image at lower dosage can be obtained without loss of image resolution at low cost and low dose.

1.5 Remainder of the Thesis


This thesis analyses the application of multiscale wavelet transform methods for denoising the low-dose computed tomography (CT) image. After formulating the problem in this chapter, the remainder of the thesis is structured as follows: Chapter 2 gives the reader an overview of literature review. The objective of the literature review is to study and analyze various methods to denoise low-dose CT image using wavelet transform. Chapter 3 discusses the history of CT generation in brief. It also gives idea about advancement in the CT generation technique. Four generations are discussed briefly. Chapter 4 briefs about the various types of noises and artifacts produced in CT images. Chapter 5 discusses various conventional spatial domain based smoothing/denoising filtering techniques and Fourier transform. It also explains the unsuitability of FT for non-stationary signals. Chapter 6 gives detail theory of the multiscale wavelet transform. It also briefs about short time Fourier transform and problem associated with it. Proposed algorithm to denoise the low-dose CT image using multiscale wavelet transform has been discussed in chapter 7. It also briefs about various quality assessment parameters that are used in the thesis to compare input low-dose noisy CT image and output denoised CT image. Chapter 8 analyzes the effectiveness and efficiency of wavelet based denoising methods and compares the results obtained using wavelet based denoising methods with that obtained using conventional denoising methods.

Chapter 1

Introduction

Finally, chapter 9 concludes the thesis with a summary of findings and several possible future directions of the research in this field of wavelet based low-dose CT image denoising.

You might also like