Robust Algorithm for Measuring Noise in Computed Tomography Examinations
Loading...
Links to Files
Permanent Link
Author/Creator
Author/Creator ORCID
Date
2020-01-20
Type of Work
Department
Computer Science and Electrical Engineering
Program
Computer Science
Citation of Original Publication
Rights
Distribution Rights granted to UMBC by the author.
This item may be protected under Title 17 of the U.S. Copyright Law. It is made available by UMBC for non-commercial research and education. For permission to publish or reproduce, please see http://aok.lib.umbc.edu/specoll/repro.php or contact Special Collections at speccoll(at)umbc.edu
This item may be protected under Title 17 of the U.S. Copyright Law. It is made available by UMBC for non-commercial research and education. For permission to publish or reproduce, please see http://aok.lib.umbc.edu/specoll/repro.php or contact Special Collections at speccoll(at)umbc.edu
Subjects
Abstract
In Computed Tomography, the image quality plays an important role in clinical diagnosis. There are many factors that contribute to poor image quality, one such factor are noise, contrast, and artifacts. Even though increased dose radiation improves the image quality, repeated examination on such dose causes harm to the patients due to the radiation exposure risk. On the other hand, low radiation causes noise thereby reducing the visibility of low contrast objects in the images which is not adequate for diagnosis purposes. Currently, there are not many optimal noise measurement and prediction methods to assess the diagnostic quality of the image in Real time at clinical practices. Previous methods in radiology physics and clinical practice measure noise by calculating the standard deviation of pixels in the homogeneous Regions Of Interests (ROIs) in the CT scan image such as soft tissue and air area outside the patient's scan. There are limitations with these methods due to its simplicity. For example, the method fails when there is no homogeneous air outside the patient to measure the noise. This work proposes a robust algorithm that uses computer vision techniques to detect the flat/smooth regions and statistical approaches to measure noise in those smooth regions of the CT scan image. We also predict noise by training on machine learning models such as linear regression, Random forest with CT examination parameters such as Slice thickness, Radiation dose, Patient size and reconstruction kernel. We are using multiple public CT for the analysis and measurement of noise. Finally, we compare the results of the measurements of noise with previous methodologies.