2D Synthetic Data Sets From Images for Deep Learning in Computed Tomography

M.S.M., Yusoff and R., Sulaiman and S., Kamarudin and F., Ramli (2021) 2D Synthetic Data Sets From Images for Deep Learning in Computed Tomography. Design Engineering, 2021 (4). 158 -168. ISSN 0011-9342

[img] PDF
Synthetic1.pdf

Download (176kB)
Official URL: http://www.thedesignengineering.com/index.php/DE/a...

Abstract

The study investigated computer-generated synthetic data setsfor use in reconstructing images in computed tomography. The data sets of computed tomography were in the format of forward projection. The present study focused on transforming a normal image data set to those of a forward projection data set. The method used to transform the normal image data set to a forward projection data set was Radon transformation. Forward projection data sets are also called sinogram data sets or the output of computed tomography data acquisition systems. The experiment was conducted using both grayscale and colour images. The time taken for the completion of the image reconstruction was measured and recorded in a table. The images from the synthetic data sets of computed tomography were plotted side by side with the original image. The synthetic data sets using forward projection data sets could be transformed back to their original image form through the application of inverse Radon transformation. This technique is known as backwards projection.

Item Type: Article
Uncontrolled Keywords: Tomography, synthetics dataset, phantom, projection, attenuation, simulation, UNIMAS, University, Borneo, Malaysia, Sarawak, Kuching, Samarahan, IPTA, education, Universiti Malaysia Sarawak
Subjects: Z Bibliography. Library Science. Information Resources > Z665 Library Science. Information Science
Z Bibliography. Library Science. Information Resources > Z665 Library Science. Information Science
Divisions: Academic Faculties, Institutes and Centres > Faculty of Computer Science and Information Technology
Depositing User: Ramli Fatihah
Date Deposited: 28 Jul 2021 04:14
Last Modified: 28 Jul 2021 04:14
URI: http://ir.unimas.my/id/eprint/35697

Actions (For repository members only: login required)

View Item View Item