Double noise filtering in CT: pre- and post-reconstruction

• Grant number: 14/11964-3
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: September 01, 2014
• End date: May 31, 2015
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
• Principal Investigator: Denis Henrique Pinheiro Salvadeo
• Grantee: Vinicius Covre de Assis
• Host Institution: Instituto de Geociências e Ciências Exatas (IGCE). Universidade Estadual Paulista (UNESP). Campus de Rio Claro. Rio Claro , SP, Brazil

Abstract

Following the ALARA principle in Computed Tomography (CT), denoising methods are needed, whether on the projections or image domain. The former we have that the projection data are corrupted by Poisson noise. Moreover, by denoising on the projections domain or sinogram can be generate artifacts in the reconstructed image from these filtered data, despite improvement in signal-to-noise ratio. Meanwhile, recent studies suggest that, whether by invoking the Central Limit Theorem or empirically, the noise on the image domain can be approximated by a Gaussian distribution. Based on these reasons, this paper proposes to perform a double noise filtering in CT, by using a state-of-art denoising methods on pre-reconstruction step such as Non Local Means for Poisson noise, reconstructing the image from these filtered projections and subsequently using a state-of-art denoising methods for Gaussian noise on the reconstructed image. Thus, we expect a better balance between noise reduction and details preservation.