Premium
Assessment of a novel compressed sensing algorithm for reconstructing phase contrast CT images of the canine prostate
Author(s) -
Montgomery James,
Zhu Zangen,
Wahid Khan
Publication year - 2013
Publication title -
the faseb journal
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.709
H-Index - 277
eISSN - 1530-6860
pISSN - 0892-6638
DOI - 10.1096/fasebj.27.1_supplement.lb45
Subject(s) - streak , compressed sensing , artificial intelligence , computer science , computer vision , image quality , projection (relational algebra) , contrast (vision) , iterative reconstruction , algorithm , wavelet transform , feature (linguistics) , image (mathematics) , wavelet , pattern recognition (psychology) , optics , linguistics , philosophy , physics
Phase contrast computed tomography (PC CT) represents a generational advance in medical and anatomical imaging with greatly improved spatial and soft tissue contrast resolution. Achieving good image quality with PC CT can require up to 4000 image projections leading to a high radiation dose and long image acquisition time. New image reconstruction methods would greatly reduce the number of projections without substantially sacrificing image quality. In sparse‐view imaging, strong streak artifacts may appear in conventionally reconstructed images, compromising image quality. Compressed sensing algorithm has shown potential to accurately recover images from highly incomplete data. The main feature of our algorithm is the use of two sparsity transforms: discrete wavelet transform and discrete gradient transform, both of which are proven to be powerful sparsity transforms. We reconstructed canine prostate images with filtered back projection and our compressed sensing algorithm using 50, 100, 120, 150, and 180 projections. Our results demonstrate that our proposed method can produce satisfactory images from only 150 projections. Research funding was provided by the Saskatchewan Health Research Fund.