Comparative Performance of Filtering Methods for Reducing Noise in Ischemic Posterior Fossa CT Images

Posterior Fossa (PF) of head Computed Tomography (CT) image appears to be one of the intricate regions as it is located in the deepest part of the brain. This appears to cause significant degradation in CT image quality due to the effect of beam hardening and bone thickness. Whilst these issues may...

Full description

Saved in:
Bibliographic Details
Published inProcedia computer science Vol. 157; pp. 55 - 63
Main Authors Muhd Suberi, Anis Azwani, Wan Zakaria, Wan Nurshazwani, Nazari, Ain, Tomari, Razali, Nik Fuad, Nik Farhan, Hj Mohd, Mohd Norzali
Format Journal Article
LanguageEnglish
Published Elsevier B.V 2019
Subjects
Computed Tomography
Image Filtering
Ischemic Stroke
Posterior Fossa
Computed Tomography
Image Filtering
Posterior Fossa
Ischemic Stroke
Online AccessGet full text

Cover

More Information
Summary:Posterior Fossa (PF) of head Computed Tomography (CT) image appears to be one of the intricate regions as it is located in the deepest part of the brain. This appears to cause significant degradation in CT image quality due to the effect of beam hardening and bone thickness. Whilst these issues may not fundamentally limit the scanning procedure, it does appear to be the contributory factors in reducing the performance of the ischemic diagnosis procedure. Thus, it is seen that image filtering is playing an important role in improving the CT image quality and effectively eliminates the influence of Poisson noise in the PF region. Therefore, this paper attempts to assess the feasibility of four different filtering methods; Anisotropic diffusion, Bilateral, Median and Wiener to eliminate noise in the CT image of PF containing ischemic. To the best of our knowledge, this is the first study to report the performance of filtering in ischemic PF. The efficacy of these four filters is evaluated in details using qualitative and quantitative metrics such as Peak Signal to Noise Ratio (PSNR), Root Mean Square Error (RMSE), Structural Similarity Index (SSIM) and processing time. The experimental works demonstrate that Bilateral filtering offers promising results in which this method can eliminate Poisson noise in CT images for ischemic PF with average PSNR, RMSE, SSIM values of 32.95, 5.7416 and 0.9749 respectively. This filter has provided the flexibility of being applicable even in cases where ischemic is presented in PF.
ISSN:1877-0509
1877-0509
DOI:10.1016/j.procs.2019.08.141