Dictionary learning based image-domain material decomposition for spectral CT.

Dictionary learning based image-domain material decomposition for spectral CT. Phys Med Biol. 2020 Jul 21;: Authors: Wu W, Yu H, Chen P, Luo F, Liu F, Wang Q, Zhu Y, Zhang Y, Feng J, Yu H Abstract The potential huge advantage of spectral computed tomography (CT) is that it can provide accurate material identification and quantitative tissue information. It is of great significance for many clinical applications, such as brain angiography, early tumor recognition, etc. To achieve more accurate material components with better image quality, a dictionary learning based image-domain material decomposition (DLIMD) is proposed for spectral CT. The procedure of DLIMD can be divided into four steps. First, spectral CT images are reconstructed from projections and then material coefficient matrix is calculated by averaging selected uniform regions of basis materials. Second, a set of image patches are extracted from the mode-1 unfolding of normalized direct inversion (DI) results to train a united dictionary by the K-SVD technique. Third, a DLIMD model is established to explore the redundant similarities from the decomposed material images. Fourth, more constraints (i.e., volume conservation and the bounds of each pixel within material maps) are integrated into the DLIMD model. The numerical mouse, physical phantom and preclinical experiments are performed to evaluate the performance of the proposed DLIMD in material decomposition accuracy, m...
Source: Physics in Medicine and Biology - Category: Physics Authors: Tags: Phys Med Biol Source Type: research