Gamma regularization based reconstruction for low dose CT. (25th August 2015)

• Record Type: Journal Article
• Authors: Zhang, Junfeng; Chen, Yang; Hu, Yining; Luo, Limin; Shu, Huazhong; Li, Bicao; Liu, Jin; Coatrieux, Jean-Louis
• Is Part Of: Physics in medicine & biology. Volume 60:Number 17(2015:Sep.)
• Page Start: 6901
• Page End: 6921
• Publication Date: 2015-08-25
• Subjects: low dose computerized tomography (LDCT) -- gamma regularization -- weighted least square (WLS); Biophysics -- Periodicals; Medical physics -- Periodicals; 610.153
• DOI: 10.1088/0031-9155/60/17/6901
• Languages: English
• ISSNs: 0031-9155

Abstract

Abstract: Reducing the radiation in computerized tomography is today a major concern in radiology. Low dose computerized tomography (LDCT) offers a sound way to deal with this problem. However, more severe noise in the reconstructed CT images is observed under low dose scan protocols (e.g. lowered tube current or voltage values). In this paper we propose a Gamma regularization based algorithm for LDCT image reconstruction. This solution is flexible and provides a good balance between the regularizations based on l 0 -norm and l 1 -norm. We evaluate the proposed approach using the projection data from simulated phantoms and scanned Catphan phantoms. Qualitative and quantitative results show that the Gamma regularization based reconstruction can perform better in both edge-preserving and noise suppression when compared with other norms.